Description
It would be very useful to store some information about containerids that are used to execute a particular query. It would allow logs for a single component to be retrieved, but also may allow better integration between workunits and the azure (or other cloud provider) information for a job - e.g. events for a container, costs?
I think we need to store something like:
containerid
jobid - if run as part of a job
component/subcomponent. e.g. thor40.slave8
local/spanid - for integrating with open telemetry?
We should also record something on a graph (subgraph?) compile etc. to associate this container id. Possibly whenever there is an Start event recorded in the stats it should have an associated or maybe a full id.
Needs more discussion because that doesn't provide a good way for the worker nodes to be associated with the graph/subgraph, and there could also be multiple spanids for a give containerid (e.g. if a new spanid is generated for each subgraph executed).