Skip to content

redis: report client calls, errors, and latency - #1242

Draft
NVGreg wants to merge 2 commits into
envoyproxy:mainfrom
NVGreg:feat/redis-operation-metrics
Draft

NVGreg wants to merge 2 commits into
envoyproxy:mainfrom
NVGreg:feat/redis-operation-metrics

Conversation

@NVGreg

@NVGreg NVGreg commented Sep 15, 2026

Copy link
Copy Markdown

Problem

Connection counts cannot show how much work the Redis client is doing or whether its calls are slow, failing, or still pending.

Change

Add fixed-label counters, an in-flight gauge, and a latency histogram for commands, pipeline groups, and startup checks. Classify returned errors without exposing keys, arguments, or error text.

Things to note

These measure synchronous client calls and attempted actions, not server command counts or background work. The histogram exports seconds. Redis and Prometheus race tests pass, including the real server metrics scope, typed errors, cancellation, blocked reporting, and both exporter unit settings. Focused vet and formatting checks pass.

Record explicit Redis client calls, outcomes, attempted actions, active calls, and duration with fixed metric scopes. Map the actual server-rooted Redis scope to Prometheus counters, gauges, and a seconds histogram, and test the constructor-to-exporter path.

Signed-off-by: Gregory Giecold <ggiecold@nvidia.com>
Signed-off-by: Gregory Giecold <ggiecold@nvidia.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant