- Renames placeholders `:orgId` to `:instanceId` in
`--app.metrics-graph`
- Adjusts frontend client to use the provided instanceId by its host
application (Weave Cloud)
When the probe is run with --probe.kubernetes.role=cluster, disable
host reporting and endpoint reporting since it duplicates what is
being done by the per-host probe.
When the probe.cri is enabled the CRI probe will be used to gather
the container information via the CRI API. For now only the basic
information is included in the generated report, those that we can get
via the CRI ListContainersRequest.
This option gives a crude way to view the raw probe data as json in
the container logs, so that you can check exactly what it would have
sent.
We stub out the PipeClient interface with a dummy implementation in
this mode.
Limiting env vars, docker&k8s labels, and weave net connection entries
to 20 is problematic because
- the truncation is arbitrary - there is a good chance that if you
care about a specific entry it won't be there
- the truncation is not consistent - different entries get truncated
at different times
- some of the rendering logic depends on specific labels, for example
namespace filtering of containers depends on the
`io.kubernetes.pod.namespace` label.
In practice, there should never be a huge number of labels, or Weave
Net connection entries. So there is no need to truncate them.
That leaves env vars. These are of limited use, so we now omit them by
default. If they are included they are included in full, so they are
actually useful.
Fixes#3127
If we have the node-name then we can save work requesting the list of
local pods from kubelet. And the best way to get the node name is via
the "downwards api", i.e. put this in the Daemonset Yaml:
- name: KUBERNETES_NODENAME
valueFrom:
fieldRef:
apiVersion: v1
fieldPath: spec.nodeName
scope-app:
* Adds `-app.metrics-graph` cli flag for configuring the base url to
use for graph links; supports :orgID and :query placeholders
* Assigns query URLs to existing metrics and appends empty metrics if missing
scope-ui:
* Extends <CloudFeature /> with option alwaysShow
* Adds <CloudLink /> to simplify routing when in cloud vs not in cloud
* Links metric graphs in the ui's node details view for all k8s
toplogies and containers so far
* Tracks metric graph click in mixpanel `scope.node.metric.click`
* Uses percentages and MB for CPU/Memory urls
* Passes timetravel timestamp to cortex in deeplink
The figure is based on our experience of running scope probes in
production systems. Below this we see a lot of "conntrack
stderr:WARNING: We have hit ENOBUFS! We are losing events."
Changed default for flag `-app.docker` to use the DOCKER_* env variables
instead of hardcoded /var/run/docker.sock; uses docker's default if
no DOCKER_HOST defined, for both probe and app.
Fixes#1975
