Bryan Boreham

Single node down when joining and leaving will cause total outage

5

If you have 3 ingesters, and replication factor 3, then a rolling update will give this error: at least 3 live ingesters required, could only find 2 This is because...

Document how to drive Ruler and AlertManager config

9

Best I could find is https://cortexmetrics.io/docs/guides/alertmanager-configuration/, which doesn't cover what changed in #3888 There are several issues which boil down to "it isn't documented": #3401, #3395, #3148

http-echo output is different to that shown

2

As of https://github.com/solsson/docker-http-echo/commit/dc2982e7d147ec4ddb962a0d0b48f380b219d9e0 the `http-echo` container no longer starts up with the `npm info` messages, so https://github.com/aws-samples/aws-workshop-for-kubernetes/blob/897d164bef18449e4c5a0266f62d57ce14481775/calico/readme.adoc#L241 is out of date. Specifically it no longer says: ``` npm info it...

probe: translate CRI container state

4

Use values consistent with what we use for Docker. This avoids seeing all containers as 'stopped', as noted at #3880.

Replace Scope metrics timeseries with Prometheus metrics

3

Make the sparkline charts in the UI source their data from a Prometheus instance that has suitable labels*. Also, if possible, make the "glass half full" overlay on nodes come...

After #3822 the merging of active controls is still a bit off - when you press 'cordon' it doesn't update for a few seconds, then you get both cordon and...

Add a button just like the ones added by #3589, but for Hosts. Now that #3822 has blazed the trail this should be straightforward. Noting that the code is quite...

I added some more logging to `app/controls.go`: ```diff @@ -95,9 +96,10 @@ func handleProbeWS(cr ControlRouter) CtxHandlerFunc { respondWith(ctx, w, http.StatusBadRequest, err) return } + log.Infof("Registered probe %s (%d)", probeID, id)...

[WIP] Probe: Better mapping of NATted connections

1

Now that we get NAT information via netlink we have a flag to distinguish `SRC_NAT` from `DST_NAT`, so we can do a better job of mapping connections. See extensive comment...

probe: avoid reporting the container ID as its hostname

2

If hostname isn't set on a container it defaults to a random hex number, which we don't want to show in the UI.