fix: litellm -> grafana
This commit is contained in:
@@ -195,6 +195,47 @@ in {
|
||||
};
|
||||
annotations.description = "A Kestra flow execution ended with FAILED, WARNING, or KILLED.";
|
||||
})
|
||||
(mkAlertRule {
|
||||
uid = "az_litellm_down";
|
||||
title = "LiteLLM metrics unreachable";
|
||||
expr = ''up{job="litellm"}'';
|
||||
for = "2m";
|
||||
evaluatorType = "lt";
|
||||
evaluatorParams = [1 0];
|
||||
labels = {
|
||||
service = "litellm";
|
||||
severity = "critical";
|
||||
};
|
||||
annotations.description = "Prometheus cannot scrape LiteLLM /metrics on AZ-CLD-1 over Netbird.";
|
||||
})
|
||||
(mkAlertRule {
|
||||
uid = "az_litellm_proxy_errors";
|
||||
title = "LiteLLM proxy errors";
|
||||
expr = ''sum(increase(litellm_proxy_failed_requests_metric_total{job="litellm"}[5m]))'';
|
||||
for = "1m";
|
||||
noDataState = "OK";
|
||||
evaluatorType = "gt";
|
||||
evaluatorParams = [0 0];
|
||||
labels = {
|
||||
service = "litellm";
|
||||
severity = "warning";
|
||||
};
|
||||
annotations.description = "LiteLLM reported one or more failed proxy responses in the last 5 minutes.";
|
||||
})
|
||||
(mkAlertRule {
|
||||
uid = "az_litellm_latency_high";
|
||||
title = "LiteLLM latency high";
|
||||
expr = ''histogram_quantile(0.95, sum(rate(litellm_request_total_latency_metric_bucket{job="litellm"}[5m])) by (le))'';
|
||||
for = "10m";
|
||||
noDataState = "OK";
|
||||
evaluatorType = "gt";
|
||||
evaluatorParams = [30 0];
|
||||
labels = {
|
||||
service = "litellm";
|
||||
severity = "warning";
|
||||
};
|
||||
annotations.description = "LiteLLM p95 total request latency has been above 30 seconds for 10 minutes.";
|
||||
})
|
||||
(mkAlertRule {
|
||||
uid = "az_http_probe_failed";
|
||||
title = "HTTP probe failed";
|
||||
|
||||
Reference in New Issue
Block a user