Daemon monitoring + Dockerfile #913
Conversation
src/http-api/routes/debug.js
Outdated
| const api = server.select('API') | ||
| const gauge = new client.Gauge({ name: 'number_of_peers', help: 'the_number_of_currently_connected_peers' }) | ||
|
|
||
| setInterval(() => { |
There was a problem hiding this comment.
Not sure about this approach. Guess we have two options, 1) (which is implemented here) is to once every X seconds, check the number of peers and save it or 2) when a request happens at /debug/metrics/prometheus, then check the number of peers and respond.
@diasdavid what you think?
|
Missing:
|
src/http-api/index.js
Outdated
| // Setup debug metrics collection | ||
| const collectDefaultMetrics = prometheusClient.collectDefaultMetrics | ||
| collectDefaultMetrics({ timeout: 5000 }) | ||
| prometheusGcStats(prometheusClient.register)() |
There was a problem hiding this comment.
Does this mean that every js-ipfs node will "phone home"?
There was a problem hiding this comment.
@diasdavid "phone home" as in send the data somewhere? No, it'll only be in the local node, available to collect when hitting the endpoint. The data is not sent somewhere.
There was a problem hiding this comment.
got it. Still, we should not add stuff that is not necessary by default. Can these be in a separate file that doesn't get loaded unless a env var is set?
There was a problem hiding this comment.
Sure, that makes sense, although it makes the usage different from go-ipfs who exposes it by default. But that makes sense, I'll move the dependencies to optionalDependencies as well then.
src/http-api/routes/debug.js
Outdated
| const api = server.select('API') | ||
| const gauge = new client.Gauge({ name: 'number_of_peers', help: 'the_number_of_currently_connected_peers' }) | ||
|
|
||
| setInterval(() => { |
|
@victorbjelkholm how close are we from this? Any update? |
|
@diasdavid just resolved a conflict, once CI is passing, is ready to be merged 👍 |
|
rebase master onto this branch. it will solve the issues you are seeing. |
For basic metrics about running process
victorb
force-pushed
the
daemon-monitoring
branch
from
7eb8369
to
5abb83d
Compare
Sets up collectors which exposes metrics at
$API/debug/metrics/prometheus(as in go-ipfs) that can be consumed by Prometheus.Ref: #804