Understanding Redis stats in use by Orchestrator Cluster

[soneill437]

I have an Orchestrator cluster setup and it is talking to the UiPath HAA Redis cluster fine. I’ve just setup a TIG Stack (Telegraf, Influxdb, Grafana) to collect and display stats including Redis stats. I’m noticing something interesting. Right now my Orchestrator cluster really isn’t doing much - no Robots currently.

Both Orchestrators are connected to all three Redis machines (via the Web.config) file. My Grafana graphs are interesting on one box. That box seems have to a constant growing total Redis connections. This is from the first Redis box I installed:

While the other two boxes are more stationary comparatively. Same graph from one of the other Redis nodes:

Another graph I think is showing what the first box is doing within Redis and it seems to be related to the “hgetall” Redis command:

What is interesting is this is the only Redis machine running that command in the Redis cluster … within these graphs that is.

I’m new to Redis and curious if anyone knows what this is related to ?

Since the Redis documentation says this: Redis HGETALL command is used to get all the fields and values of the hash stored at the key. Is this related to something Orchestrator is doing ?

And if so, why would only a single machine be getting a majority of the “hgetall” commands ?

[soneill437]

Without fully understanding it, pretty sure this is related to the very first Redis machine installed is the Master of the cluster. The other machines are slaves/replicas and more specifically they are Read-Only replicas from going through the Redis documentation. Additional Telegraf/Grafana charts confirm the read-only replica status.