fix: Grafana Dashboard definitions #983
Conversation
| @@ -38,14 +38,15 @@ | |||
| { | |||
| "datasource": { | |||
| "type": "prometheus", | |||
| "uid": "PBFA97CFB590B2093" | |||
| "uid": "Prometheus" | |||
There was a problem hiding this comment.
This is to align the UID set in the prometheus.yml file in the configuration, instead of the custom UID set on Dashboard creation
| }, | ||
| "fieldConfig": { | ||
| "defaults": { | ||
| "color": { | ||
| "mode": "palette-classic" | ||
| }, | ||
| "custom": { | ||
| "axisBorderShow": false, |
There was a problem hiding this comment.
These are new fields set as default per Grafana
| "expr": "histogram_quantile(0.9, max by(le, method) (increase(ponder_database_method_duration_bucket{service=\"admin\"}[$__rate_interval])))", | ||
| "expr": "histogram_quantile(0.9, max by(le, method) (increase(ponder_database_method_duration_bucket{service=\"internal\"}[$__rate_interval])))", |
There was a problem hiding this comment.
Service admin is now internal for setup, migrateSyncStore, publish, updateFinalizedCheckpoint, and createIndexes
| "editorMode": "code", | ||
| "exemplar": false, | ||
| "expr": "histogram_quantile(0.95, ponder_realtime_rpc_request_duration_bucket)", | ||
| "expr": "histogram_quantile(0.95, max by(le, method) (increase(ponder_rpc_request_duration_bucket[$__rate_interval])))", |
There was a problem hiding this comment.
Renamed field in the /metrics endpoint
| "targets": [ | ||
| { | ||
| "datasource": { | ||
| "type": "prometheus", | ||
| "uid": "Prometheus" | ||
| }, | ||
| "editorMode": "builder", | ||
| "expr": "ponder_realtime_latest_block_number", | ||
| "expr": "ponder_realtime_latest_block_number{network=\"$networks\"}", |
There was a problem hiding this comment.
Added filter parameter (re-used the one already set on ponder_realtime_latest_block_timestamp)
| @@ -1040,7 +1085,7 @@ | |||
| }, | |||
| "editorMode": "code", | |||
| "exemplar": false, | |||
| "expr": "histogram_quantile(0.5, ponder_historical_rpc_request_duration_bucket)", | |||
There was a problem hiding this comment.
historical is no longer available on the /metrics endpoint, what to do? delete?
These are under a Historical section on the Dashboard, and they're already available for the Live section
| @@ -1114,7 +1164,7 @@ | |||
| }, | |||
| "editorMode": "code", | |||
| "exemplar": false, | |||
| "expr": "histogram_quantile(0.95, ponder_historical_rpc_request_duration_bucket)", | |||
| "expr": "histogram_quantile(0.95, ponder_rpc_request_duration_bucket)", | |||
There was a problem hiding this comment.
All other fields are renamed fields still available and exposed by Ponder, and default Grafana updates
|
Here's an overview of the dashboard with fixed graph's
|
Wanted to have better overview of my Ponder instances and discovered that there was a Prometheus and Grafana setup available (there seems not to be mentioned anywhere in the README or the documentation, TO-DO?) and successfully managed to setup instances and data ingestion on Railway (after a quite long fail-and-retry process).
Noticed a bunch of graph's were out-of-date and using deprecated and/or renamed fields exposed on the
/metricsendpoint. Documentation make it well-aware that this is not the final API, but I found it quite helpful to have a bunch of definitions to quickstart everything, so I went ahead and updated all graph's as required.There is still a bunch of things to address before this can be merged, I've added comment's where input is appreciated.