FreeswitchSessionsCritical #
High sessions usage on {{ $labels.instance }}: {{ $value | printf “%.2f”}}%
Alert Rule
alert: FreeswitchSessionsCritical
annotations:
description: |-
High sessions usage on {{ $labels.instance }}: {{ $value | printf "%.2f"}}%
VALUE = {{ $value }}
LABELS = {{ $labels }}
runbook: https://srerun.github.io/prometheus-alerts/runbooks/znerol-freeswitch-exporter/freeswitchsessionscritical/
summary: Freeswitch Sessions Critical (instance {{ $labels.instance }})
expr: (freeswitch_session_active * 100 / freeswitch_session_limit) > 90
for: 5m
labels:
severity: critical
Meaning #
The FreeswitchSessionsCritical alert is triggered when the percentage of active FreeSWITCH sessions exceeds 90% of the total session limit. This indicates that the system is experiencing high usage and may be at risk of exhausting available resources.
Impact #
If left unaddressed, high session usage can lead to:
- Degraded system performance
- Increased latency and response times
- Potential crashes or failures of the FreeSWITCH server
- Impact on overall system stability and reliability
Diagnosis #
To diagnose the root cause of the high session usage, follow these steps:
- Check the FreeSWITCH server logs for any errors or anomalies
- Verify that the session limit is set correctly and adjust if necessary
- Investigate any recent changes to the system or configuration that may be contributing to the high usage
- Review the FreeSWITCH debug logs to identify any specific issues or bottlenecks
- Check the system resource utilization (CPU, memory, etc.) to ensure it is within acceptable limits
Mitigation #
To mitigate the high session usage, follow these steps:
- Increase the session limit if possible
- Identify and terminate any idle or unnecessary sessions
- Optimize system configuration and tuning to improve performance and reduce resource utilization
- Implement rate limiting or other traffic control measures to prevent sudden spikes in session usage
- Consider load balancing or distributing traffic across multiple FreeSWITCH servers to reduce the load on individual instances