Our AWS Lambda invocation counts vary quite a lot over time, as our users are primarily in a narrow band of timezones. So, setting up a single threshold for alerting based on error counts is not very meaningful, due to the spikes in invocations. How can I configure an alert based on the ratio of invocations to errors (e.g. alert on > 5% error rate)?
Related topics
| Topic | Replies | Views | Activity | |
|---|---|---|---|---|
| How to calculate error percentage in the alert | 4 | 112 | February 25, 2026 | |
| AWS CloudWatch Metrics alert rule - Alert on any error except from exclusion list | 1 | 46 | March 27, 2026 | |
| Question: Loki Alerting – Inconsistent Data in Alert Notifications | 0 | 42 | June 5, 2025 | |
| Disk used percentage alert | 0 | 494 | October 26, 2023 | |
| Grafana Alert fails while fetching query results | 2 | 4415 | October 31, 2023 |