Http related metrics coming from Quarkus are different from Vertx

[ShubhamRwt]

Hi, I have some question related to the quarkus micrometer metrics.

1. Why the http related metrics coming from quarkus different then the http related related metrics coming through vertx if Quarkus is based upon vertx?

2. We get this metric http_server_bytes_read_sum from Quarkus but with vertx we get http_server_bytes_read_total. Is totalandsum` the same thing?

3. We get this metric http_server_connections_seconds_active_count from quarkus. What does seconds mean here. Is it similar to the vertx metrics http_server_requests_total?

FYI - @ppatierno

[quarkus-bot[bot]]

/cc @ebullient (metrics), @jmartisk (metrics)

[ebullient]

Quarkus integrates micrometer much differently than vertx does. So while yes, Quarkus uses vertx as its underlying request engine, its integration with micrometer is very different and more direct.

The suffixes are provided by the Micrometer Prometheus Meter Registry.

• http_server_bytes_read_sum should be the same as http_server_bytes_read_total.
• http_server_connections_seconds_active_count is a long task timer. It shows how many seconds the longest running connection has been active for.
• There should also be http_server_requests_total (and related) that present metrics for completed requests.

[ShubhamRwt]

These are all the http_server_requests_seconds related metrics

[ShubhamRwt]

There are no metrics as just http_server_requests. The strimzi_bridge_ is a prefix we have added.

[ShubhamRwt]

@ebullient ^^^

[ebullient]

• http_server_requests_seconds_* -- Yes. that is the set for completed server requests.
• http_server_connections_seconds_active_* are long-task timers

The former measures completed requests (successful or not), the latter measures active/open connections.

[ShubhamRwt]

Thanks @ebullient