Scaling: Difference between revisions

From VoIPmonitor.org
(Review: oprava wiki syntaxe (kategorie, tabulky, formátování, PHP syntaxhighlight))
(Add link to Blog in See Also)
 
(14 intermediate revisions by 2 users not shown)
Line 2: Line 2:
[[Category:Administration]]
[[Category:Administration]]


This guide provides a comprehensive overview of performance tuning and scaling for VoIPmonitor. It covers the three primary system bottlenecks and offers practical, expert-level advice for optimizing your deployment for high traffic loads.
This guide covers performance tuning for high-traffic VoIPmonitor deployments, addressing the three primary system bottlenecks.


== Understanding Performance Bottlenecks ==
== Understanding Performance Bottlenecks ==
A VoIPmonitor deployment's maximum capacity is determined by three potential bottlenecks. Identifying and addressing the correct one is key to achieving high performance.
 
A VoIPmonitor deployment's capacity is limited by three potential bottlenecks:


<kroki lang="plantuml">
<kroki lang="plantuml">
Line 11: Line 12:
skinparam shadowing false
skinparam shadowing false
skinparam defaultFontName Arial
skinparam defaultFontName Arial
skinparam rectangle {
  BorderColor #4A90E2
  BackgroundColor #FFFFFF
}


title VoIPmonitor Performance Bottlenecks
title VoIPmonitor Performance Bottlenecks
Line 22: Line 19:
rectangle "RTP/SIP\nProcessing" as PROC #E6FFE6
rectangle "RTP/SIP\nProcessing" as PROC #E6FFE6
rectangle "PCAP Files\nStorage" as DISK #FFF3E6
rectangle "PCAP Files\nStorage" as DISK #FFF3E6
database "MySQL/MariaDB\nDatabase" as DB #E6E6FF
database "MySQL/MariaDB" as DB #E6E6FF


NIC -right-> T0 : "1. CPU\nBottleneck"
NIC -right-> T0 : "1. CPU"
T0 -right-> PROC
T0 -right-> PROC
PROC -down-> DISK : "2. I/O\nBottleneck"
PROC -down-> DISK : "2. I/O"
PROC -right-> DB : "3. Database\nBottleneck"
PROC -right-> DB : "3. Database"


note bottom of T0
note bottom of T0
   Monitor: t0CPU in syslog
   Monitor: t0CPU
   Limit: Single CPU core
   Limit: 1 CPU core
end note
end note


note bottom of DISK
note bottom of DISK
   Monitor: iostat, ioping
   Monitor: iostat
   Solution: SSD, TAR archives
   Solution: SSD, TAR
end note
end note


note bottom of DB
note bottom of DB
   Monitor: SQLq in syslog
   Monitor: SQLq
   Solution: Partitioning, tuning
   Solution: RAM, tuning
end note
end note
@enduml
@enduml
</kroki>
</kroki>


The three bottlenecks are:
{| class="wikitable"
# '''Packet Capturing (CPU & Network Stack):''' The ability of a single CPU core to read packets from the network interface. This is often the first limit encountered.
|-
# '''Disk I/O (Storage):''' The speed at which the sensor can write PCAP files to disk. Critical when call recording is enabled.
! Bottleneck !! Description !! Monitor
# '''Database Performance (MySQL/MariaDB):''' The rate at which the database can ingest CDRs and serve data to the GUI.
|-
| '''1. Packet Capture''' || Single CPU core reading packets from NIC || <code>t0CPU</code> in syslog
|-
| '''2. Disk I/O''' || Writing PCAP files to storage || '''<code>IO[C%]</code> in syslog''' (v2026.01.3+), <code>iostat</code>
|-
| '''3. Database''' || CDR ingestion and GUI queries || <code>SQLq</code> in syslog
|}


On a modern, well-tuned server (e.g., 24-core Xeon, 10Gbit NIC), a single VoIPmonitor instance can handle up to '''10,000 concurrent calls''' with full RTP analysis and recording, or over '''60,000 concurrent calls''' with SIP-only analysis.
'''Capacity:''' A modern server (24-core Xeon, 10Gbit NIC) can handle '''~10,000 concurrent calls''' with full RTP recording, or '''60,000+''' with SIP-only analysis.


== Optimizing Packet Capturing (CPU & Network) ==
== Optimizing Packet Capture (CPU & Network) ==
The most performance-critical task is the initial packet capture, handled by a single, highly optimized thread (t0). If this thread's CPU usage (<code>t0CPU</code> in logs) approaches 100%, you are hitting the capture limit.


=== Use a Modern Linux Kernel & VoIPmonitor Build ===
The packet capture thread (t0) runs on a single CPU core. If <code>t0CPU</code> approaches 100%, you've hit the capture limit.
Modern Linux kernels (3.2+) and VoIPmonitor builds include '''TPACKET_V3''' support, a high-speed packet capture mechanism. This is the single most important factor for high performance.


'''Recommendation:''' Always use a recent Linux distribution (AlmaLinux, Rocky Linux, or Debian) and the latest VoIPmonitor static binary. With this combination, a standard Intel 10Gbit NIC can often handle up to 2 Gbit/s of VoIP traffic without special drivers.
With a modern kernel and VoIPmonitor build, a standard Intel 10Gbit NIC handles up to 3 Gbit/s VoIP traffic without special drivers and almost full 10Gbit rate with [[DPDK]]


=== Network Stack & Driver Tuning ===
=== Threading (Automatic) ===
For high-traffic environments (>500 Mbit/s), fine-tuning the network driver and kernel parameters is essential.


==== NIC Ring Buffer ====
Since version 2023.11, VoIPmonitor uses <code>threading_expanded=yes</code> by default, which automatically spawns threads based on CPU load. '''No manual threading configuration is needed.'''
The ring buffer is a queue between the network card driver and VoIPmonitor. A larger buffer prevents packet loss during short CPU usage spikes.


<syntaxhighlight lang="bash">
For very high traffic (≥1500 Mbit/s), set:
# Check maximum size
<syntaxhighlight lang="ini">
ethtool -g eth0
threading_expanded = high_traffic
</syntaxhighlight>


# Set to maximum (e.g., 16384)
See [[Sniffer_configuration#Threading_Model|Threading Model]] for details.
ethtool -G eth0 rx 16384
</syntaxhighlight>


==== Interrupt Coalescing ====
=== NIC Tuning (>500 Mbit/s) ===
This setting batches multiple hardware interrupts into one, reducing CPU overhead.


<syntaxhighlight lang="bash">
<syntaxhighlight lang="bash">
# Increase ring buffer (prevents packet loss during CPU spikes)
ethtool -g eth0                  # Check max size
ethtool -G eth0 rx 16384        # Set to max
# Enable interrupt coalescing (reduces CPU overhead)
ethtool -C eth0 rx-usecs 1022
ethtool -C eth0 rx-usecs 1022
</syntaxhighlight>
</syntaxhighlight>


==== Applying Settings Persistently ====
'''Persistent settings''' (Debian/Ubuntu <code>/etc/network/interfaces</code>):
To make these settings permanent, add them to your network configuration. For Debian/Ubuntu using <code>/etc/network/interfaces</code>:
 
<syntaxhighlight lang="ini">
<syntaxhighlight lang="ini">
auto eth0
auto eth0
Line 94: Line 94:
</syntaxhighlight>
</syntaxhighlight>


Note: Modern systems using NetworkManager or systemd-networkd require different configuration methods.
=== Configuration Optimizations ===


==== Configuration-Level Optimizations ====
{| class="wikitable"
Before investing in kernel-bypass solutions, ensure your <code>voipmonitor.conf</code> is optimized for performance. Several configuration parameters can significantly reduce CPU load and improve packet capture efficiency.
|-
! Parameter !! Purpose !! Recommendation
|-
| <code>interface_ip_filter</code> || IP-based filtering || More efficient than BPF <code>filter</code>
|-
| <code>pcap_dump_writethreads_max</code> || Compression threads || Set to CPU core count
|-
| <code>jitterbuffer_f1/f2/adapt</code> || Jitter simulation || Keep <code>f2=yes</code>, disable f1 and adapt to save CPU while keeping MOS
|}


;Use interface_ip_filter Instead of filter
<syntaxhighlight lang="ini">
:If you need to filter by IP address or subnet, use <code>interface_ip_filter</code> instead of the general BPF <code>filter</code> option. The <code>interface_ip_filter</code> directive is more efficient and reduces CPU overhead compared to complex BPF filters.
# /etc/voipmonitor.conf


<syntaxhighlight lang="ini">
# Efficient IP filtering (replaces BPF filter)
# More efficient IP-based filtering
interface_ip_filter = 192.168.0.0/24
interface_ip_filter = 192.168.0.0/24
interface_ip_filter = 10.0.0.0/8
interface_ip_filter = 10.0.0.0/8


# Less efficient BPF filtering (avoid if possible)
# Compression scaling
# filter = udp and (host 192.168.0.0/24 or host 10.0.0.0/8)
</syntaxhighlight>
 
{{Note|See [[Sniffer_configuration]] for the complete reference and description of <code>interface_ip_filter</code> (Interface Selection section).}}
 
;Optimize PCAP Compression Threads
:For systems with high call recording rates, PCAP compression can become CPU-intensive. VoIPmonitor can automatically scale compression threads.
 
<syntaxhighlight lang="ini">
# /etc/voipmonitor.conf
# Initial compression threads (auto-scales based on load)
pcap_dump_writethreads = 1
pcap_dump_writethreads = 1
# Maximum compression threads (adjust based on CPU cores)
pcap_dump_writethreads_max = 32
pcap_dump_writethreads_max = 32
# Asynchronous PCAP writing (enabled by default)
pcap_dump_asyncwrite = yes
pcap_dump_asyncwrite = yes
</syntaxhighlight>
</syntaxhighlight>


{{Tip|Set <code>pcap_dump_writethreads_max</code> to the number of CPU cores available for best performance on multi-core systems. Monitor <code>t0CPU</code> to ensure compression threads are not competing with the capture thread.}}
{{Note|1=Recommended: <code>jitterbuffer_f1=no</code>, <code>jitterbuffer_f2=yes</code>, <code>jitterbuffer_adapt=no</code>. This saves CPU while preserving MOS-F2 metrics. Only disable f2 if you don't need quality monitoring at all.}}
 
;Adjust Jitterbuffer Settings Based on Traffic Patterns
:Jitterbuffer simulation adds CPU overhead. For production environments with stable networks, consider adjusting jitterbuffer settings to balance accuracy with performance.


<syntaxhighlight lang="ini">
=== Kernel-Bypass Solutions ===
# /etc/voipmonitor.conf
# Fixed 50ms jitterbuffer (default: yes)
jitterbuffer_f1 = yes


# Fixed 200ms jitterbuffer (default: yes)
For extreme loads, bypass the kernel network stack entirely:
jitterbuffer_f2 = yes
 
# Adaptive jitterbuffer up to 500ms (default: yes)
jitterbuffer_adapt = yes
</syntaxhighlight>
 
{{Warning|Disabling jitterbuffer analysis reduces CPU load but removes MOS and jitter quality metrics from CDRs. Only disable if you do not require voice quality monitoring.}}
 
=== Advanced Kernel-Bypass Solutions ===
If kernel and driver tuning are insufficient, you can offload the capture process entirely by bypassing the kernel's network stack.


{| class="wikitable"
{| class="wikitable"
Line 156: Line 132:
| '''[[DPDK]]''' || Open-source || ~70% || Multi-gigabit on commodity hardware
| '''[[DPDK]]''' || Open-source || ~70% || Multi-gigabit on commodity hardware
|-
|-
| '''PF_RING ZC/DNA''' || Commercial || 90% → 20% || High-volume enterprise
| '''PF_RING ZC''' || Commercial || 90% → 20% || High-volume enterprise
|-
|-
| '''Napatech SmartNICs''' || Hardware || <3% at 10 Gbit/s || Extreme performance requirements
| '''[[Napatech|Napatech SmartNICs]]''' || Hardware || <3% at 10Gbit/s || Extreme performance
|}
|}
;DPDK (Data Plane Development Kit)
:A set of libraries and drivers for fast packet processing. VoIPmonitor can leverage DPDK to read packets directly from the network card, completely bypassing the kernel. See [[DPDK|DPDK guide]] for details.
;PF_RING ZC/DNA
:A commercial software driver from ntop.org that dramatically reduces CPU load by bypassing the kernel.
;Napatech SmartNICs
:Specialized hardware acceleration cards that deliver packets with near-zero CPU overhead.


== Optimizing Disk I/O ==
== Optimizing Disk I/O ==
VoIPmonitor's modern storage engine is highly optimized to minimize random disk access, which is the primary cause of I/O bottlenecks.


=== VoIPmonitor Storage Strategy ===
=== VoIPmonitor Storage Strategy ===
Instead of writing a separate PCAP file for each call (which causes massive I/O load), VoIPmonitor groups all calls starting within the same minute into a single compressed <code>.tar</code> archive. This changes the I/O pattern from thousands of small, random writes to a few large, sequential writes, reducing IOPS by a factor of 10 or more.


'''Typical capacity:''' A standard 7200 RPM SATA drive can handle up to 2,000 concurrent calls with full recording.
VoIPmonitor groups all calls starting within the same minute into a single compressed <code>.tar</code> archive. This changes thousands of random writes into few sequential writes, reducing IOPS by 10x+.
 
'''Typical capacity:''' 7200 RPM SATA handles ~2,000 concurrent calls with full recording.


=== Filesystem Tuning (ext4) ===
=== Filesystem Tuning (ext4) ===
For the spool directory (<code>/var/spool/voipmonitor</code>), using an optimized ext4 filesystem can improve performance.


<syntaxhighlight lang="bash">
<syntaxhighlight lang="bash">
# Format partition without a journal (use with caution, requires battery-backed RAID controller)
# Format without journal (requires battery-backed RAID)
mke2fs -t ext4 -O ^has_journal /dev/sda2
mke2fs -t ext4 -O ^has_journal /dev/sda2
</syntaxhighlight>


# Add to /etc/fstab for optimal performance
<syntaxhighlight lang="ini">
/dev/sda2   /var/spool/voipmonitor  ext4   errors=remount-ro,noatime,data=writeback,barrier=0 0 0
# /etc/fstab
/dev/sda2 /var/spool/voipmonitor  ext4 errors=remount-ro,noatime,data=writeback,barrier=0 0 0
</syntaxhighlight>
</syntaxhighlight>


{{Warning|Disabling the journal removes protection against filesystem corruption after crashes. Only use this with a battery-backed RAID controller.}}
{{Warning|1=Disabling journal removes crash protection. Only use with battery-backed RAID controller (BBU).}}


=== RAID Controller Cache Policy ===
=== RAID Controller ===
A misconfigured RAID controller is a common bottleneck. For database and spool workloads, the cache policy should be set to '''WriteBack''', not WriteThrough. This applies for RPM disks, not fast SSDs.


'''Requirements:'''
Set cache policy to '''WriteBack''' (not WriteThrough). Requires healthy BBU. Commands vary by vendor (<code>megacli</code>, <code>ssacli</code>, <code>perccli</code>).
* A healthy Battery Backup Unit (BBU) is required
* Specific commands vary by vendor (<code>megacli</code>, <code>ssacli</code>, <code>perccli</code>)
* Refer to vendor documentation for LSI, HP, and Dell controllers


== Optimizing Database Performance (MySQL/MariaDB) ==
== Optimizing Database Performance ==
A well-tuned database is critical for both data ingestion from the sensor and GUI responsiveness.


{{Note|For extreme scenarios (4,000+ concurrent calls, UI lag, high SQL queue, or 1000+ CDRs/sec), see [[High-Performance_VoIPmonitor_and_MySQL_Setup_Manual]] for specialized configurations including innodb_flush_log_at_trx_commit=0, hourly partitioning, and centralized writer architecture.}}
=== Memory Configuration ===


=== Memory Configuration ===
The most critical parameter is <code>innodb_buffer_pool_size</code>.
The most critical database parameter is <code>innodb_buffer_pool_size</code>, which defines how much memory InnoDB uses to cache data and indexes.


{{Warning|On servers running both VoIPmonitor and MySQL, setting <code>innodb_buffer_pool_size</code> too high causes OOM (Out of Memory) killer events, resulting in CDR delays, crashes, and instability. See [[Sniffer_troubleshooting#Check_for_OOM_.28Out_of_Memory.29_Issues|OOM Troubleshooting]] for details.}}
{{Warning|1=Setting too high causes OOM killer events, CDR delays, and crashes. See [[Sniffer_troubleshooting#Check_for_OOM_.28Out_of_Memory.29_Issues|OOM Troubleshooting]].}}


==== Buffer Pool Sizing ====
'''Buffer Pool Sizing:'''


{| class="wikitable"
{| class="wikitable"
|-
|-
! Server Type !! Calculation !! Example (32GB RAM)
! Server Type !! Formula !! Example (32GB RAM)
|-
|-
| '''Shared''' (VoIPmonitor + MySQL) || (Total RAM - VoIPmonitor - OS overhead) / 2 || 14GB
| '''Shared''' (VoIPmonitor + MySQL) || (Total RAM - VoIPmonitor - OS) / 2 || 14GB
|-
|-
| '''Dedicated''' MySQL server || 50-70% of total RAM || 20-22GB
| '''Dedicated''' MySQL server || 50-70% of total RAM || 20-22GB
|}
|}


For shared servers, use this formula:
'''RAM Recommendations:'''
<syntaxhighlight lang="text">
innodb_buffer_pool_size = (Total RAM - VoIPmonitor memory - OS overhead - safety margin) / 2
 
Example for a 32GB server:
- Total RAM: 32GB
- VoIPmonitor process memory: ~2GB (check with ps aux)
- OS + other services overhead: ~2GB
- Available for buffer pool: 28GB
- Recommended innodb_buffer_pool_size = 14G
</syntaxhighlight>


==== RAM Recommendations ====
{| class="wikitable"
{| class="wikitable"
|-
|-
! Deployment Size !! Minimum RAM !! Recommended RAM
! Deployment Size !! Minimum !! Recommended
|-
|-
| Small (<500 concurrent calls) || 8GB || 16GB
| Small (<500 calls) || 8GB || 16GB
|-
|-
| Medium (500-2000 calls) || 16GB || 32GB
| Medium (500-2000) || 16GB || 32GB
|-
|-
| Large (>2000 calls) || 32GB || 64GB+
| Large (>2000) || 32GB || 64GB+
|}
|}


==== Disable Graphical Desktop ====
=== Key MySQL Parameters ===
A graphical desktop environment consumes 1-2GB of RAM unnecessarily. VoIPmonitor is managed through a web interface and does not require a desktop.
 
<syntaxhighlight lang="bash">
# Disable display manager
systemctl stop gdm          # Ubuntu/Debian with GDM
systemctl disable gdm
 
# Set default to multi-user (no GUI)
systemctl set-default multi-user.target
 
# Verify memory freed
free -h
</syntaxhighlight>
 
=== Other Key Parameters ===


<syntaxhighlight lang="ini">
<syntaxhighlight lang="ini">
# /etc/mysql/my.cnf or /etc/mysql/mariadb.conf.d/50-server.cnf
# /etc/mysql/my.cnf or mariadb.conf.d/50-server.cnf
 
[mysqld]
[mysqld]
# Buffer pool size (calculate per above)
innodb_buffer_pool_size = 14G
innodb_buffer_pool_size = 14G
 
innodb_flush_log_at_trx_commit = 2  # Faster, minimal data loss risk
# Flush logs to OS cache, write to disk once per second (faster, minimal data loss risk)
innodb_file_per_table = 1           # Essential for partitioning
innodb_flush_log_at_trx_commit = 2
innodb_compression_algorithm = lz4 # MariaDB only
 
# Store each table in its own file (essential for partitioning)
innodb_file_per_table = 1
 
# LZ4 compression for modern MariaDB
innodb_compression_algorithm = lz4
</syntaxhighlight>
</syntaxhighlight>
{{Warning|For deployments with 4,000+ concurrent calls experiencing UI unresponsiveness, database queue growth (SQLq), or extremely high CDR insertion rates (1000+ CDRs/sec), the intermediate settings above may not be sufficient. See [[High-Performance_VoIPmonitor_and_MySQL_Setup_Manual]] for extreme performance configurations optimized for 10,000+ concurrent calls, including innodb_flush_log_at_trx_commit=0, hourly partitioning, and centralized writer architecture.}}


=== Slow Query Log ===
=== Slow Query Log ===


The MySQL slow query log can consume significant memory and disk I/O on high-traffic systems. If you are experiencing high memory utilization alerts or performance issues with the database server, consider adjusting or disabling the slow query log.
The slow query log can consume significant memory. Consider disabling on high-traffic systems:
 
{{Warning|Disabling the slow query log removes the ability to analyze slow queries for performance optimization. Only disable it temporarily or if you are certain you do not need it.}}


<syntaxhighlight lang="ini">
<syntaxhighlight lang="ini">
# /etc/mysql/my.cnf or /etc/my.cnf.d/mysql-server.cnf
[mysqld]
[mysqld]
# Disable slow query log (set to 1 to enable)
slow_query_log = 0
slow_query_log = 0
 
# Or increase threshold: long_query_time = 600
# Alternative: Increase threshold to only log extremely slow queries (e.g., 600 seconds = 10 minutes)
long_query_time = 600
</syntaxhighlight>
</syntaxhighlight>


After changing MySQL configuration, restart the database and dependent services:
=== Database Partitioning ===
 
<syntaxhighlight lang="bash">
# Restart MySQL/MariaDB
systemctl restart mariadb  # or mysql


# Restart VoIPmonitor sniffer (depends on database)
VoIPmonitor automatically partitions large tables (like <code>cdr</code>) by day. This is enabled by default and '''highly recommended'''.
systemctl restart voipmonitor
</syntaxhighlight>


=== Troubleshooting: Connection Refused Errors ===
See [[Data_Cleaning#Database_Cleaning_.28CDR_Retention.29|Database Partitioning]] for details.


If the VoIPmonitor GUI or sniffer experiences crashes with "Can't connect to MySQL server" or "Connection refused" errors, verify that MariaDB is configured with adequate memory.
=== Troubleshooting: Connection Refused ===


'''Symptoms:'''
'''Symptoms:''' GUI crashes, "Connection refused" errors, intermittent issues during peak volumes.
* GUI server crashes repeatedly
* MariaDB connection refused errors in logs
* MariaDB service appears to be running (`systemctl status mariadb` shows active)
* Issues occur intermittently during peak call volumes


'''Cause:'''
'''Cause:''' <code>innodb_buffer_pool_size</code> too low (default 128M is insufficient).
When <code>innodb_buffer_pool_size</code> is set too low (e.g., default 128M or 256M), MariaDB may run out of memory for the InnoDB buffer pool during high call volumes or complex database queries. This can cause connection failures even while the database service remains technically running.


'''Solution:'''
'''Solution:''' Increase to 6G+ based on available RAM:
Edit the MariaDB configuration file (path varies by distribution):
 
<syntaxhighlight lang="bash">
# Common configuration file locations:
# Debian/Ubuntu: /etc/mysql/mariadb.conf.d/50-server.cnf
# RHEL/CentOS: /etc/my.cnf.d/mariadb-server.cnf
# Generic: /etc/mysql/my.cnf
 
# Edit the file and increase innodb_buffer_pool_size
nano /etc/mysql/mariadb.conf.d/50-server.cnf
</syntaxhighlight>
 
Set a higher buffer pool size (example for a server with 8-16GB RAM):


<syntaxhighlight lang="ini">
<syntaxhighlight lang="ini">
[mysqld]
[mysqld]
# Increase from default 128M to 6G or higher, depending on available RAM
innodb_buffer_pool_size = 6G
innodb_buffer_pool_size = 6G
</syntaxhighlight>
</syntaxhighlight>
Restart MariaDB to apply the change:


<syntaxhighlight lang="bash">
<syntaxhighlight lang="bash">
Line 347: Line 239:
</syntaxhighlight>
</syntaxhighlight>


{{Tip|Use the [[#Buffer Pool Sizing|Buffer Pool Sizing formula]] above to calculate the optimal value for your server. For most production deployments with 500-2000 concurrent calls, values between 6G and 14G are typical.}}
== Component Separation (Multi-Host Architecture) ==
 
=== Database Partitioning ===
VoIPmonitor automatically splits large tables (like <code>cdr</code>) into daily partitions. This is enabled by default and '''highly recommended'''.
 
'''Benefits:'''
* Massively improves GUI query performance (only relevant partitions are scanned)
* Allows instant deletion of old data by dropping partitions (thousands of times faster than DELETE)
 
See [[Data_Cleaning#The_Modern_Method:_Partitioning_.28Recommended.29|Database Partitioning]] for configuration details.
 
== Scaling Through Component Separation ==
 
For deployments requiring high availability and independent scaling of different workloads, VoIPmonitor's three core components (MySQL database, Sensor, and GUI) can be deployed on separate dedicated hosts. This architecture allows each component to be scaled based on its specific resource requirements.
 
=== When to Use Component Separation ===
 
Consider separating components when:
* You need to scale packet capture without affecting GUI responsiveness
* Database queries are becoming a bottleneck during peak hours
* You have multiple teams managing different aspects of the system (DBAs, network engineers, web administrators)
* You want to implement different high availability strategies for each component
* Your traffic volume exceeds 5,000-10,000 concurrent calls


=== Three-Host Architecture ===
For deployments exceeding 5,000-10,000 concurrent calls, separate VoIPmonitor components onto dedicated hosts.


In this model, each core VoIPmonitor component runs on a dedicated server:
=== Architecture Overview ===


{| class="wikitable"
{| class="wikitable"
Line 379: Line 249:
! Host !! Component !! Primary Resources !! Scaling Strategy
! Host !! Component !! Primary Resources !! Scaling Strategy
|-
|-
| '''Host 1: MySQL Database''' || MariaDB/MySQL server || RAM (50-70% for buffer pool), fast I/O (SSD/NVMe) || Add RAM, upgrade storage, add read replicas for reporting
| '''Host 1''' || MySQL Database || RAM, fast SSD || Add RAM, read replicas
|-
|-
| '''Host 2: Sensor(s)''' || Packet capture and analysis || Single CPU core (t0 thread), network I/O || Add more sensors for distributed capture, use DPDK/PF_RING
| '''Host 2''' || Sensor(s) || CPU (t0 thread), network || DPDK/PF_RING, more sensors
|-
|-
| '''Host 3: GUI''' || Web interface (PHP + Apache/Nginx) || CPU for database queries, network bandwidth || Horizontal scaling with load balancer, caching
| '''Host 3''' || GUI || CPU, network || Load balancer, caching
|}
|}


=== Configuration Steps ===
=== Configuration ===
 
==== Step 1: Deploy Dedicated MySQL Server ====
 
Configure the database server with proper MySQL/MariaDB tuning for VoIPmonitor. Use the dedicated server buffer pool sizing from [[#Memory_Configuration|Memory Configuration]].


'''MySQL Server:'''
<syntaxhighlight lang="ini">
<syntaxhighlight lang="ini">
# /etc/mysql/my.cnf on MySQL Host
# /etc/mysql/my.cnf
[mysqld]
[mysqld]
bind-address = 0.0.0.0 # Allow remote connections
bind-address = 0.0.0.0
innodb_buffer_pool_size = 50G  # 50-70% of total RAM for dedicated DB server
innodb_buffer_pool_size = 50G  # 50-70% RAM for dedicated server
innodb_flush_log_at_trx_commit = 2
innodb_file_per_table = 1
innodb_compression_algorithm = lz4
</syntaxhighlight>
</syntaxhighlight>
Grant access for remote connections:


<syntaxhighlight lang="sql">
<syntaxhighlight lang="sql">
CREATE USER 'voipmonitor'@'%' IDENTIFIED BY 'strong_password';
CREATE USER 'voipmonitor'@'%' IDENTIFIED BY 'strong_password';
GRANT ALL PRIVILEGES ON voipmonitor.* TO 'voipmonitor'@'%';
GRANT ALL PRIVILEGES ON voipmonitor.* TO 'voipmonitor'@'%';
FLUSH PRIVILEGES;
</syntaxhighlight>
</syntaxhighlight>


==== Step 2: Deploy Sensor(s) ====
'''Sensor:'''
 
Install the VoIPmonitor sensor on one or more dedicated capture servers. Configure the sensor to connect to the remote MySQL database.
 
<syntaxhighlight lang="ini">
<syntaxhighlight lang="ini">
# /etc/voipmonitor.conf on Sensor Host
# /etc/voipmonitor.conf
[general]
id_sensor = 1
id_sensor = 1 # Unique per sensor
mysqlhost = mysql.server.ip
# Remote MySQL connection
mysqlhost = mysql.server.ip # IP of dedicated MySQL server
mysqldb = voipmonitor
mysqldb = voipmonitor
mysqluser = voipmonitor
mysqlusername = voipmonitor
mysqlpassword = strong_password
mysqlpassword = strong_password
# Capture settings
interface = eth0
sipport = 5060,5061
# Performance tuning
pcap_dump_writethreads = 1
pcap_dump_writethreads_max = 32
</syntaxhighlight>
==== Step 3: Deploy Dedicated GUI ====
Install the VoIPmonitor GUI on a separate web server and configure it to connect to the remote MySQL database.
<syntaxhighlight lang="bash">
# Extract GUI package
tar -xzf voipmonitor-gui-latest.tar.gz -C /var/www/html/
# Edit GUI configuration
/var/www/html/voipmonitor/config/system_configuration.php
</syntaxhighlight>
Configure database connection through the GUI interface:
# Open <nowiki>http://gui.server.ip/voipmonitor</nowiki> in a browser
# Navigate to '''Settings > System Configuration > Database'''
# Configure:
# * Host: mysql.server.ip
# * User: voipmonitor
# * Password: strong_password
# * Database: voipmonitor
Alternatively, edit the configuration file directly:
<syntaxhighlight lang="php">
// /var/www/html/voipmonitor/config/system_configuration.php
define('DB_HOST', 'mysql.server.ip');
define('DB_USER', 'voipmonitor');
define('DB_PASS', 'strong_password');
define('DB_NAME', 'voipmonitor');
</syntaxhighlight>
</syntaxhighlight>


==== Network Connectivity Requirements ====
'''GUI:''' Configure via Settings > System Configuration > Database, or edit <code>config/system_configuration.php</code>.
 
Ensure firewall rules allow:


'''Firewall Rules:'''
{| class="wikitable"
{| class="wikitable"
|-
|-
! Source !! Destination !! Port !! Purpose
! Source !! Destination !! Port !! Purpose
|-
|-
| Sensor host || MySQL server || 3306 || Write CDRs to database
| Sensor || MySQL || 3306 || CDR writes
|-
|-
| GUI host || MySQL server || 3306 || Read/Write for queries and configuration
| GUI || MySQL || 3306 || Queries
|-
|-
| User browsers || GUI host || 80, 443 || Access web interface
| GUI || Sensor(s) || 5029 || PCAP retrieval
|-
|-
| (Optional) GUI host || Sensor host(s) || 5029 || Retrieve PCAP files (if using local processing mode)
| Users || GUI || 80, 443 || Web access
|}
|}


=== Advantages of Component Separation ===
{{Note|1=Component separation can be combined with [[Sniffer_distributed_architecture|Client-Server mode]] for multi-site deployments.}}
 
* '''Independent Scaling''' - Each component can be upgraded independently based on load
* '''Resource Optimization''' - Database servers need more RAM, sensors need fast CPUs, GUI needs network and query performance
* '''Isolation''' - Issues on one component do not affect others (e.g., GUI upgrade does not stop packet capture)
* '''Team Specialization''' - DBAs manage database, network engineers manage sensors, web team manages GUI
* '''High Availability Options''' - Different HA strategies for each component:
** MySQL: Galera cluster for multi-master replication
** Sensors: Client-server mode with failover
** GUI: Load balancer with multiple web servers


=== Comparison with Distributed Client-Server Mode ===
== Monitoring Performance ==


The three-host component separation architecture differs from [[Sniffer_distributed_architecture|Client-Server mode]]:
VoIPmonitor logs performance metrics every 10 seconds to syslog. Key metrics to watch:


{| class="wikitable"
{| class="wikitable"
|-
|-
! Architecture !! Scenario !! Best For
! Metric !! Warning Sign !! Bottleneck Type
|-
| <code>t0CPU</code> || >90% || CPU (packet capture limit)
|-
| '''<code>IO[C%]</code>''' || '''C ≥ 80% (WARN), C ≥ 95% (DISK_SAT)''' || '''Disk I/O (v2026.01.3+)'''
|-
| <code>heap[A&#124;B&#124;C]</code> || A >50% || I/O or CPU (buffer filling)
|-
|-
| '''Component Separation (this section)''' || 3 dedicated hosts: MySQL + GUI + Sensor(s) || Single data center, independent scaling needs, centralized database
| <code>SQLq</code> || Growing || Database
|-
|-
| '''Client-Server Mode''' || Central server + remote sensors forwarding packets || Multi-site deployments, geographic distribution, low-resource remote sensors
| <code>comp</code> || Maxed out || I/O (compression waiting for disk)
|}
|}
Both models can be combined: Use component separation at the central site, while remote sites use client-server mode to forward data.
=== Monitoring Component Health ===
Monitor each component separately:
* '''MySQL Server:''' Check <code>innodb_buffer_pool_size</code>, query throughput, slow queries
* '''Sensor(s):''' Monitor <code>t0CPU</code>, packet drops, SQL queue depth
* '''GUI Server:''' Monitor web server load, PHP-FPM statistics, page response times
Use the [[Sniffer_troubleshooting|Sniffer Troubleshooting]] guide for sensor issues and database performance tools for MySQL tuning.
== Monitoring Live Performance ==
VoIPmonitor logs detailed performance metrics every 10 seconds to syslog.


<syntaxhighlight lang="bash">
<syntaxhighlight lang="bash">
# Debian/Ubuntu
# Monitor in real-time
tail -f /var/log/syslog | grep voipmonitor
journalctl -u voipmonitor -f
</syntaxhighlight>


# CentOS/RHEL
'''Main article: [[Syslog_Status_Line]]''' - Complete reference for all metrics with detailed explanations and troubleshooting guidance.
tail -f /var/log/messages | grep voipmonitor
</syntaxhighlight>


=== Understanding the Log Output ===
'''For bottleneck diagnosis:''' See [[Sniffer_troubleshooting#Diagnose:_I.2FO_vs_CPU_Bottleneck|I/O vs CPU Bottleneck Diagnosis]] for step-by-step diagnostic procedure using syslog metrics and Linux tools.
Sample log line:
<syntaxhighlight lang="text">
voipmonitor[15567]: calls[315][355] PS[C:4 S:29/29 R:6354 A:6484] SQLq[0] heap[0|0|0] comp[54] [12.6Mb/s] t0CPU[5.2%] ... RSS/VSZ[323|752]MB
</syntaxhighlight>


{| class="wikitable"
== See Also ==
|-
! Metric !! Description !! Warning Threshold
|-
| <code>calls[X][Y]</code> || X = active calls, Y = total calls in memory || -
|-
| <code>SQLq[C]</code> || SQL queries waiting to be sent to database || Growing consistently = DB bottleneck
|-
| <code>heap[A{{!}}B{{!}}C]</code> || Memory usage % for internal buffers || A = 100% → packet drops
|-
| <code>t0CPU[X%]</code> || '''Main packet capture thread CPU usage''' || >90-95% = capture limit reached
|-
| <code>RSS/VSZ[X{{!}}Y]MB</code> || Resident/Virtual memory usage || RSS growing = memory leak
|}


=== Performance Diagrams ===
* [[Sniffer_troubleshooting]] - Troubleshooting including OOM issues
* [[Data_Cleaning]] - Database and spool retention
* [[Sniffer_configuration]] - Complete configuration reference
* [[DPDK]] - DPDK setup guide
* [[Sniffer_distributed_architecture]] - Client-Server mode
* [[Blog#January_2026:_Enhanced_Sensor_Monitoring_.26_Disk_I.2FO_Analytics|Blog: January 2026 Release Notes]] - Detailed overview of new Disk I/O monitoring and GUI improvements


The following diagrams illustrate the difference between standard kernel packet capture and optimized solutions:


[[File:kernelstandarddiagram.png|thumb|center|600px|Standard kernel packet capture path - packets traverse multiple kernel layers before reaching VoIPmonitor]]


[[File:ntop.png|thumb|center|600px|PF_RING/DPDK bypass mode - packets are delivered directly to VoIPmonitor, bypassing the kernel network stack]]


== See Also ==
* [[Sniffer_troubleshooting]] - Troubleshooting guide including OOM issues
* [[Data_Cleaning]] - Database and spool retention configuration
* [[Sniffer_configuration]] - Complete configuration reference
* [[DPDK]] - DPDK setup guide
* [[IO_Measurement]] - Disk I/O benchmarking tools


== AI Summary for RAG ==
== AI Summary for RAG ==
'''Summary:''' Expert guide to scaling VoIPmonitor for high-traffic environments. Covers three main bottlenecks: (1) Packet Capturing - optimized via TPACKET_V3, NIC tuning with ethtool (ring buffer, interrupt coalescing), configuration-level optimizations (interface_ip_filter more efficient than BPF filter, pcap_dump_writethreads for compression thread tuning, jitterbuffer settings for CPU/performance balance), and kernel-bypass solutions (DPDK, PF_RING, Napatech); (2) Disk I/O - VoIPmonitor uses TAR-based storage to reduce IOPS, with ext4 tuning and RAID WriteBack cache; (3) Database - critical innodb_buffer_pool_size tuning with formula for shared servers: (Total RAM - VoIPmonitor - OS overhead) / 2. For 32GB shared server, recommend 14GB buffer pool. Dedicated servers can use 50-70% of RAM. Covers slow query log as a memory/I/O consumer and disabling it for memory optimization. Covers partitioning benefits and syslog monitoring (t0CPU, SQLq, heap metrics). TROUBLESHOOTING: GUI or sniffer crashes with "MariaDB connection refused" errors indicate innodb_buffer_pool_size is set too low (e.g., default 128M or 256M). This causes MariaDB to run out of memory for the InnoDB buffer pool during high call volumes, resulting in connection failures even while the service remains active. Solution: Edit MariaDB config file (typically /etc/mysql/mariadb.conf.d/50-server.cnf or /etc/my.cnf.d/mariadb-server.cnf) and increase innodb_buffer_pool_size to 6G or higher based on available RAM. For most production deployments with 500-2000 concurrent calls, values between 6G and 14G are typical. Restart MariaDB with systemctl restart mariadb after changing the configuration. For extreme scenarios (4,000+ concurrent calls, UI lag, unresponsive GUI, high SQL queue), see High-Performance_VoIPmonitor_and_MySQL_Setup_Manual for specialized configurations including innodb_flush_log_at_trx_commit=0, hourly partitioning, centralized writer architecture, RTP thread tuning (rtpthreads, rtpthreads_start), and MySQL optimization settings (innodb_thread_concurrency, innodb_io_capacity, innodb_flush_method=O_DIRECT).


COMPONENT SEPARATION ARCHITECTURE: For deployments requiring independent scaling, VoIPmonitor's three core components (MySQL database, Sensor, and GUI) can be deployed on separate dedicated hosts. This architecture is recommended when needing to scale packet capture without affecting GUI responsiveness, database queries are becoming a bottleneck, multiple teams manage different aspects (DBAs, network engineers, web administrators), traffic volume exceeds 5,000-10,000 concurrent calls, or different HA strategies are needed. Three-Host Architecture: Host 1 is MySQL Database (needs 50-70% RAM for buffer pool, fast SSD/NVMe I/O, scale by adding RAM or read replicas, consider Galera cluster for HA); Host 2 is Sensor(s) (needs fast CPU single core for t0 thread and network I/O, scale by adding more sensors for distributed capture or use DPDK/PF_RING); Host 3 is GUI web interface (needs CPU for database queries and network bandwidth, can be horizontally scaled with load balancer and caching, use multiple web servers with load balancer for HA). Configuration: MySQL server with bind-address=0.0.0.0 and remote user access via CREATE USER 'voipmonitor'@'%' with GRANT ALL PRIVILEGES. Sensor configured with mysqlhost pointing to MySQL server IP. GUI configured via Settings > System Configuration > Database or by editing config/system_configuration.php to set DB_HOST, DB_USER, DB_PASS, DB_NAME. Network connectivity requires firewall rules: Sensor→MySQL (port 3306), GUI→MySQL (port 3306), Users→GUI (ports 80/443), GUI→Sensor(s) (port 5029 for PCAP retrieval if using local processing). Advantages: independent scaling, resource optimization (database needs RAM, sensors need CPU, GUI needs network), isolation (issues on one component don't affect others), team specialization, different HA strategies per component. Comparison with Client-Server Mode: Component Separation (3 dedicated hosts MySQL+GUI+Sensor) is best for single data center and independent scaling needs while Client-Server Mode (central server+remote sensors forwarding packets) is best for multi-site and geographic distribution with low-resource remote sensors. Both models can be combined. Monitor components separately: MySQL (innodb_buffer_pool_size, query throughput, slow queries), Sensors (t0CPU, packet drops, SQL queue), GUI (web server load, PHP-FPM, response times).
<!-- This section is for AI/RAG systems. Do not edit manually. -->
 
=== Summary ===
Performance tuning guide for high-traffic VoIPmonitor deployments. Covers three main bottlenecks: CPU (t0 packet capture thread, single-core limit), Disk I/O (PCAP storage), and Database (MySQL/MariaDB). Threading is automatic since 2023.11 via threading_expanded=yes (use high_traffic for ≥1500 Mbit/s). NIC tuning: ethtool ring buffer and interrupt coalescing. CPU optimization: interface_ip_filter instead of BPF, jitterbuffer_f2=yes with f1/adapt disabled. Kernel bypass solutions: DPDK (~70% CPU reduction), PF_RING ZC, Napatech SmartNICs (<3% CPU at 10Gbit). Disk I/O: TAR archives reduce IOPS 10x, ext4 tuning (noatime, writeback), RAID WriteBack cache with BBU. Database: innodb_buffer_pool_size (50-70% RAM for dedicated server), innodb_flush_log_at_trx_commit=2, partitioning. Multi-host architecture for >5000-10000 concurrent calls separating MySQL, sensors, and GUI.


'''Keywords:''' scaling, performance tuning, bottleneck, t0CPU, TPACKET_V3, DPDK, PF_RING, ethtool, ring buffer, interface_ip_filter, BPF filter, pcap_dump_writethreads, jitterbuffer, jitterbuffer_f1, jitterbuffer_f2, jitterbuffer_adapt, compression threads, PCAP async write, innodb_buffer_pool_size, OOM killer, shared server memory, database partitioning, SQLq monitoring, slow query log, slow_query_log, long_query_time, UI lag, unresponsive GUI, high performance, 4000 concurrent calls, 5000 concurrent calls, innodb_flush_log_at_trx_commit=0, hourly partitioning, rtpthreads, rtpthreads_start, RTP threads, innodb_io_capacity, innodb_thread_concurrency, innodb_flush_method, extreme performance, High-Performance Manual, component separation, three host architecture, dedicated MySQL server, dedicated GUI server, remote database, mysqlhost configuration, GUI remote database, independent scaling, high availability, Galera cluster, load balancer, DB_HOST, DB_USER, DB_PASS, DB_NAME, bind-address, firewall ports, GUI settings database configuration, connection refused, MariaDB connection refused, GUI crashes, database connection errors, mariadb.conf.d/50-server.cnf, MySQL crash, MariaDB crash, can't connect to MySQL
=== Keywords ===
scaling, performance, tuning, optimization, high traffic, bottleneck, CPU, t0CPU, t0 thread, single-core, disk I/O, storage, database, MySQL, MariaDB, threading_expanded, high_traffic, NIC tuning, ethtool, ring buffer, interrupt coalescing, interface_ip_filter, jitterbuffer, DPDK, PF_RING, Napatech, kernel bypass, TAR archive, ext4, noatime, writeback, RAID, WriteBack cache, BBU, innodb_buffer_pool_size, innodb_flush_log_at_trx_commit, partitioning, multi-host, component separation, concurrent calls, capacity, 10000 calls, heap, SQLq, compression threads, pcap_dump_writethreads


'''Key Questions:'''
=== Key Questions ===
* How do I scale VoIPmonitor for thousands of concurrent calls?
* How to tune VoIPmonitor for high traffic?
* What are the main performance bottlenecks in VoIPmonitor?
* How many concurrent calls can VoIPmonitor handle?
* How do I fix high t0CPU usage?
* What are the main performance bottlenecks?
* What is DPDK and when should I use it?
* How to optimize CPU usage for packet capture?
* How do I calculate innodb_buffer_pool_size for a shared server?
* What is threading_expanded and when to use high_traffic?
* What happens if innodb_buffer_pool_size is set too high?
* How to tune NIC for VoIPmonitor?
* What causes "MariaDB connection refused" errors in the GUI or sniffer?
* How to reduce CPU with jitterbuffer settings?
* How do I fix GUI crashes due to MariaDB connection refused errors?
* What is DPDK and when to use it?
* How do I troubleshoot database connection issues with MariaDB?
* How to optimize disk I/O for PCAP storage?
* How do I interpret the performance metrics in syslog?
* How to tune ext4 filesystem for VoIPmonitor?
* Should I use a dedicated database server for VoIPmonitor?
* What is the recommended innodb_buffer_pool_size?
* How much RAM does a VoIPmonitor server need?
* How to configure MySQL for VoIPmonitor performance?
* How can the slow query log affect memory utilization?
* When to separate VoIPmonitor components to multiple hosts?
* How do I disable or adjust the MySQL slow query log?
* How to monitor VoIPmonitor performance metrics?
* Is interface_ip_filter more efficient than the filter option?
* What do t0CPU, heap, SQLq metrics mean?
* How do I optimize PCAP compression threads for high traffic?
* Which jitterbuffer settings affect CPU load the most?
* What configuration options reduce CPU overhead?
* How do I deploy VoIPmonitor with MySQL, GUI, and Sensor on separate servers?
* When should I use component separation architecture for VoIPmonitor?
* What is the three host architecture for VoIPmonitor scaling?
* How do I configure the GUI to connect to a remote MySQL database?
* How do I configure sensors to write to a remote MySQL server?
* What are the advantages of separating VoIPmonitor components onto different hosts?
* How do I scale the MySQL database independently from sensors and GUI?
* How do I scale the GUI independently from database and sensors?
* What is the difference between component separation and client-server mode?
* Can I combine component separation with client-server mode for VoIPmonitor?
* How do I configure firewall rules for component separation architecture?
* When should I use a dedicated GUI server for VoIPmonitor?

Latest revision as of 18:35, 23 January 2026


This guide covers performance tuning for high-traffic VoIPmonitor deployments, addressing the three primary system bottlenecks.

Understanding Performance Bottlenecks

A VoIPmonitor deployment's capacity is limited by three potential bottlenecks:

Bottleneck Description Monitor
1. Packet Capture Single CPU core reading packets from NIC t0CPU in syslog
2. Disk I/O Writing PCAP files to storage IO[C%] in syslog (v2026.01.3+), iostat
3. Database CDR ingestion and GUI queries SQLq in syslog

Capacity: A modern server (24-core Xeon, 10Gbit NIC) can handle ~10,000 concurrent calls with full RTP recording, or 60,000+ with SIP-only analysis.

Optimizing Packet Capture (CPU & Network)

The packet capture thread (t0) runs on a single CPU core. If t0CPU approaches 100%, you've hit the capture limit.

With a modern kernel and VoIPmonitor build, a standard Intel 10Gbit NIC handles up to 3 Gbit/s VoIP traffic without special drivers and almost full 10Gbit rate with DPDK

Threading (Automatic)

Since version 2023.11, VoIPmonitor uses threading_expanded=yes by default, which automatically spawns threads based on CPU load. No manual threading configuration is needed.

For very high traffic (≥1500 Mbit/s), set: 

threading_expanded = high_traffic

See Threading Model for details.

NIC Tuning (>500 Mbit/s)

# Increase ring buffer (prevents packet loss during CPU spikes)
ethtool -g eth0                  # Check max size
ethtool -G eth0 rx 16384         # Set to max

# Enable interrupt coalescing (reduces CPU overhead)
ethtool -C eth0 rx-usecs 1022

Persistent settings (Debian/Ubuntu /etc/network/interfaces): 

auto eth0
iface eth0 inet manual
    up ip link set $IFACE up
    up ip link set $IFACE promisc on
    up ethtool -G $IFACE rx 16384
    up ethtool -C $IFACE rx-usecs 1022

Configuration Optimizations

Parameter Purpose Recommendation
interface_ip_filter IP-based filtering More efficient than BPF filter
pcap_dump_writethreads_max Compression threads Set to CPU core count
jitterbuffer_f1/f2/adapt Jitter simulation Keep f2=yes, disable f1 and adapt to save CPU while keeping MOS 
# /etc/voipmonitor.conf

# Efficient IP filtering (replaces BPF filter)
interface_ip_filter = 192.168.0.0/24
interface_ip_filter = 10.0.0.0/8

# Compression scaling
pcap_dump_writethreads = 1
pcap_dump_writethreads_max = 32
pcap_dump_asyncwrite = yes

ℹ️ Note: Recommended: jitterbuffer_f1=no, jitterbuffer_f2=yes, jitterbuffer_adapt=no. This saves CPU while preserving MOS-F2 metrics. Only disable f2 if you don't need quality monitoring at all.

Kernel-Bypass Solutions

For extreme loads, bypass the kernel network stack entirely:

Solution Type CPU Reduction Use Case
DPDK Open-source ~70% Multi-gigabit on commodity hardware
PF_RING ZC Commercial 90% → 20% High-volume enterprise
Napatech SmartNICs Hardware <3% at 10Gbit/s Extreme performance

Optimizing Disk I/O

VoIPmonitor Storage Strategy

VoIPmonitor groups all calls starting within the same minute into a single compressed .tar archive. This changes thousands of random writes into few sequential writes, reducing IOPS by 10x+.

Typical capacity: 7200 RPM SATA handles ~2,000 concurrent calls with full recording.

Filesystem Tuning (ext4)

# Format without journal (requires battery-backed RAID)
mke2fs -t ext4 -O ^has_journal /dev/sda2

# /etc/fstab
/dev/sda2  /var/spool/voipmonitor  ext4  errors=remount-ro,noatime,data=writeback,barrier=0  0 0

⚠️ Warning: Disabling journal removes crash protection. Only use with battery-backed RAID controller (BBU).

RAID Controller

Set cache policy to WriteBack (not WriteThrough). Requires healthy BBU. Commands vary by vendor (megacli, ssacli, perccli).

Optimizing Database Performance

Memory Configuration

The most critical parameter is innodb_buffer_pool_size.

⚠️ Warning: Setting too high causes OOM killer events, CDR delays, and crashes. See OOM Troubleshooting.

Buffer Pool Sizing:

Server Type Formula Example (32GB RAM)
Shared (VoIPmonitor + MySQL) (Total RAM - VoIPmonitor - OS) / 2 14GB
Dedicated MySQL server 50-70% of total RAM 20-22GB

RAM Recommendations:

Deployment Size Minimum Recommended
Small (<500 calls) 8GB 16GB
Medium (500-2000) 16GB 32GB
Large (>2000) 32GB 64GB+

Key MySQL Parameters

# /etc/mysql/my.cnf or mariadb.conf.d/50-server.cnf
[mysqld]
innodb_buffer_pool_size = 14G
innodb_flush_log_at_trx_commit = 2  # Faster, minimal data loss risk
innodb_file_per_table = 1           # Essential for partitioning
innodb_compression_algorithm = lz4  # MariaDB only

Slow Query Log

The slow query log can consume significant memory. Consider disabling on high-traffic systems: 

[mysqld]
slow_query_log = 0
# Or increase threshold: long_query_time = 600

Database Partitioning

VoIPmonitor automatically partitions large tables (like cdr) by day. This is enabled by default and highly recommended.

See Database Partitioning for details.

Troubleshooting: Connection Refused

Symptoms: GUI crashes, "Connection refused" errors, intermittent issues during peak volumes.

Cause: innodb_buffer_pool_size too low (default 128M is insufficient).

Solution: Increase to 6G+ based on available RAM: 

[mysqld]
innodb_buffer_pool_size = 6G

systemctl restart mariadb

Component Separation (Multi-Host Architecture)

For deployments exceeding 5,000-10,000 concurrent calls, separate VoIPmonitor components onto dedicated hosts.

Architecture Overview

Host Component Primary Resources Scaling Strategy
Host 1 MySQL Database RAM, fast SSD Add RAM, read replicas
Host 2 Sensor(s) CPU (t0 thread), network DPDK/PF_RING, more sensors
Host 3 GUI CPU, network Load balancer, caching

Configuration

MySQL Server: 

# /etc/mysql/my.cnf
[mysqld]
bind-address = 0.0.0.0
innodb_buffer_pool_size = 50G  # 50-70% RAM for dedicated server

CREATE USER 'voipmonitor'@'%' IDENTIFIED BY 'strong_password';
GRANT ALL PRIVILEGES ON voipmonitor.* TO 'voipmonitor'@'%';

Sensor: 

# /etc/voipmonitor.conf
id_sensor = 1
mysqlhost = mysql.server.ip
mysqldb = voipmonitor
mysqlusername = voipmonitor
mysqlpassword = strong_password

GUI: Configure via Settings > System Configuration > Database, or edit config/system_configuration.php.

Firewall Rules:

Source Destination Port Purpose
Sensor MySQL 3306 CDR writes
GUI MySQL 3306 Queries
GUI Sensor(s) 5029 PCAP retrieval
Users GUI 80, 443 Web access

ℹ️ Note: Component separation can be combined with Client-Server mode for multi-site deployments.

Monitoring Performance

VoIPmonitor logs performance metrics every 10 seconds to syslog. Key metrics to watch:

Metric Warning Sign Bottleneck Type
t0CPU >90% CPU (packet capture limit)
IO[C%] C ≥ 80% (WARN), C ≥ 95% (DISK_SAT) Disk I/O (v2026.01.3+)
heap[A|B|C] A >50% I/O or CPU (buffer filling)
SQLq Growing Database
comp Maxed out I/O (compression waiting for disk) 
# Monitor in real-time
journalctl -u voipmonitor -f

Main article: Syslog_Status_Line - Complete reference for all metrics with detailed explanations and troubleshooting guidance.

For bottleneck diagnosis: See I/O vs CPU Bottleneck Diagnosis for step-by-step diagnostic procedure using syslog metrics and Linux tools.

See Also



AI Summary for RAG

Summary

Performance tuning guide for high-traffic VoIPmonitor deployments. Covers three main bottlenecks: CPU (t0 packet capture thread, single-core limit), Disk I/O (PCAP storage), and Database (MySQL/MariaDB). Threading is automatic since 2023.11 via threading_expanded=yes (use high_traffic for ≥1500 Mbit/s). NIC tuning: ethtool ring buffer and interrupt coalescing. CPU optimization: interface_ip_filter instead of BPF, jitterbuffer_f2=yes with f1/adapt disabled. Kernel bypass solutions: DPDK (~70% CPU reduction), PF_RING ZC, Napatech SmartNICs (<3% CPU at 10Gbit). Disk I/O: TAR archives reduce IOPS 10x, ext4 tuning (noatime, writeback), RAID WriteBack cache with BBU. Database: innodb_buffer_pool_size (50-70% RAM for dedicated server), innodb_flush_log_at_trx_commit=2, partitioning. Multi-host architecture for >5000-10000 concurrent calls separating MySQL, sensors, and GUI.

Keywords

scaling, performance, tuning, optimization, high traffic, bottleneck, CPU, t0CPU, t0 thread, single-core, disk I/O, storage, database, MySQL, MariaDB, threading_expanded, high_traffic, NIC tuning, ethtool, ring buffer, interrupt coalescing, interface_ip_filter, jitterbuffer, DPDK, PF_RING, Napatech, kernel bypass, TAR archive, ext4, noatime, writeback, RAID, WriteBack cache, BBU, innodb_buffer_pool_size, innodb_flush_log_at_trx_commit, partitioning, multi-host, component separation, concurrent calls, capacity, 10000 calls, heap, SQLq, compression threads, pcap_dump_writethreads

Key Questions

  • How to tune VoIPmonitor for high traffic?
  • How many concurrent calls can VoIPmonitor handle?
  • What are the main performance bottlenecks?
  • How to optimize CPU usage for packet capture?
  • What is threading_expanded and when to use high_traffic?
  • How to tune NIC for VoIPmonitor?
  • How to reduce CPU with jitterbuffer settings?
  • What is DPDK and when to use it?
  • How to optimize disk I/O for PCAP storage?
  • How to tune ext4 filesystem for VoIPmonitor?
  • What is the recommended innodb_buffer_pool_size?
  • How to configure MySQL for VoIPmonitor performance?
  • When to separate VoIPmonitor components to multiple hosts?
  • How to monitor VoIPmonitor performance metrics?
  • What do t0CPU, heap, SQLq metrics mean?
Retrieved from "https://www.voipmonitor.org/doc/index.php?title=Scaling&oldid=11113"