docs: add New Relic monitoring dashboard for forwarder.* metrics - #52
docs: add New Relic monitoring dashboard for forwarder.* metrics#52voorepreethi wants to merge 2 commits into
Conversation
Adds docs/oci-log-forwarder-metrics-dashboard-template.json — a pre-built dashboard covering all 17 forwarder.* custom metrics across both basic and advanced tiers: - Page 1 (Basic Metrics): invocations, record flow, delivery health, pipeline lag — useful at metrics_tier=basic or advanced. - Page 2 (Advanced Metrics): byte volumes, error breakdown by class, batching/performance, client cache hit rate — useful at metrics_tier=advanced. Also adds a "Monitoring dashboard" section to README.md with import instructions and a summary of each page's content. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
| } | ||
| ], | ||
| "platformOptions": { | ||
| "ignoreTimeRange": false |
There was a problem hiding this comment.
Hardcoded SINCE 1 hour ago on the "Forwarder info" table, while every other widget respects the dashboard's time-range picker,this is inconsistent — if someone changes the dashboard to view a 24-hour or 7-day window, this one table silently stays locked to the last hour.
There was a problem hiding this comment.
It overrides the time stamp from the main time picker . So won't be an issue due to this flag
| { | ||
| "name": "OCI Log Forwarder monitoring", | ||
| "description": null, | ||
| "permissions": "PUBLIC_READ_WRITE", |
There was a problem hiding this comment.
Here anyone can edit this dashboard by hand in the UI.Is this intended or we should have PUBLIC_READ_ONLY?
There was a problem hiding this comment.
If needed they can edit the dashboard to add more widgets. So added PUBLIC_READ_WRITE. Let me know if we have to move it PUBLIC_READ_ONLY
| { | ||
| "alertSeverity": "WARNING", | ||
| "value": 0 | ||
| } |
There was a problem hiding this comment.
threshold looks miscalibrated. It sets alertSeverity: WARNING at value: 0 on a widget whose first column is 'Total' invocations — that would visually flag the tile as a warning basically all the time (any nonzero traffic crosses "> 0").
There was a problem hiding this comment.
Record drop rate: dropped / received * 100
Byte delivery ratio: delivered / received * 100
Client cache hit rate: hit / total * 100
If the denominator is ever 0 (e.g. right after a fresh deploy with zero traffic, or a tenancy with a still-cold cache), NRQL division returns null rather than erroring.how is each visualization renders a null (blank vs. "0%" vs. a UI error), especially since two of these have severity thresholds attached.
…tiles The Invocations billboard's WARNING threshold bound to the Total column, which is nonzero during all normal traffic, so it flagged every healthy period as a warning. Also documents that Record drop rate, Byte delivery ratio, and Client cache hit rate render blank (NRQL null) rather than 0% when their denominator is zero, e.g. right after a fresh deploy. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
jashraf-hash
left a comment
jashraf-hash
left a comment
There was a problem hiding this comment.
I feel here it is only considering 1 hour range from current time.Is that the intended use or we want to support any date range?
|
It supports all time ranges from time picker |
3 participants
Summary
docs/oci-log-forwarder-metrics-dashboard-template.json— a pre-built New Relic dashboard template covering all 17forwarder.*custom metrics emitted by the OCI log forwarder.README.mdwith import instructions and a page-by-page summary.Dashboard pages
Page 1 — Basic Metrics (available at
metrics_tier=basicoradvanced):Page 2 — Advanced Metrics (available at
metrics_tier=advanced):Stacking
This PR is stacked on #48 (advanced metrics). It should be merged after #48 lands on
NR-601759-oci-observability.Test plan
docs/oci-log-forwarder-metrics-dashboard-template.jsoninto New Relic (Dashboards → Import dashboard) after replacingYOUR_ACCOUNT_ID.metrics_tier=basic.metrics_tier=advanced.