[RedisEnterprise] Enable use_type flag for node and proxy datastreams - #8213
Conversation
🌐 Coverage report
|
| "node": "1", | ||
| "instance": "host.docker.internal:8070", | ||
| "job": "prometheus" | ||
| }, |
There was a problem hiding this comment.
is this a coincidence that all node* metrics are gauge/value?
| "listener_total_started_res": 0, | ||
| "listener_cmd_set_max": 0, | ||
| "listener_acc_latency_max": 0 | ||
| "listener_resp2_clients_max": { |
There was a problem hiding this comment.
Same here, is this a coincidence that all proxy* metrics are counters?
There was a problem hiding this comment.
I will verify this again.
| type: object | ||
| object_type: keyword | ||
| description: Label fields | ||
| - name: redisenterprise.node.*.value |
There was a problem hiding this comment.
Assuming the metric_type mapping will be taken in separate PR.
There was a problem hiding this comment.
Yes, those will be raised separately as part of the TSDB migration.
There was a problem hiding this comment.
Assuming the metric type mappsing is: redisenterprise.node.*.value will directly map to gauge and redisenterprise.proxy.*.counter to counter.
There was a problem hiding this comment.
Yes, we do need to explicitly specify the metric_type, and I will address this in the metric_type PR.
lalit-satapathy
left a comment
There was a problem hiding this comment.
LGTM, Need to follow-up and ensure the TSDB mappings are updated accordingly.
| description: > | ||
| Redisenterprise node gauge metric | ||
|
|
||
| - name: redisenterprise.node.*.counter |
There was a problem hiding this comment.
Is this entry needed, since node is all gauge? Or is it a fallback if any nodes have counter in the future?
There was a problem hiding this comment.
I believe it's a better idea to remove it. Instead of retaining a fallback, we can update it when a new field is introduced in a new version. I will also make sure to update the tested version for consistency.
| type: object | ||
| object_type: keyword | ||
| description: Label fields | ||
| - name: redisenterprise.node.*.value |
There was a problem hiding this comment.
Assuming the metric type mappsing is: redisenterprise.node.*.value will directly map to gauge and redisenterprise.proxy.*.counter to counter.
| "listener_total_started_res": 0, | ||
| "listener_cmd_set_max": 0, | ||
| "listener_acc_latency_max": 0 | ||
| "listener_resp2_clients_max": { |
agithomas
left a comment
There was a problem hiding this comment.
Please verify if a change in the snapshot image is needed as part of the dashboard change?
|
I suppose there is |
| Redis Enterprise integration provides monitoring of [redis](https://redis.com/) cluster. Monitoring is done via prometheus exported port of redis enterprise cluster. Once a redis enterprise [cluster](https://redis.com/redis-enterprise/technology/redis-enterprise-cluster-architecture/) is installed, prometheus port is available for monitoring. The url of the host:port(8070) needs to be passed to the hosts in the settings. | ||
|
|
||
| Redis Enterpise integration is tested with redislabs/redis:5.2.2-24 version. | ||
| Redis Enterpise integration is tested with redislabs/redis:7.2.4-72 version. |
There was a problem hiding this comment.
Please add a title ## Compatibility above this line.
There aren't any major changes, just some adjustments to field names that aren't visible in the snapshot. |
Sorry, I don't understand why we need to overturn it. |
Ignore it, it can be taken up later in a separate PR and it is a part of package spec V3 issue addressal. |
|
Left one more important comment: #8213 (comment) |
|
Package redisenterprise - 0.7.0 containing this change is available at https://epr.elastic.co/search?package=redisenterprise |

Proposed commit message
This PR enables use_type flag for node and proxy datastreams of RedisEnterprise. This is required to correctly specify the metric_type mapping, as use_type enables a different layout for metrics storage, leveraging Elasticsearch types.
Checklist
changelog.ymlfile.Related issues
Screenshots