[ci] Improving error reporting and reproducibility #1195
Description
Currently in our ci we are printing links to query the logs for each node, as part of the regular reporting (both for success and fail jobs). We want to improve the ux when test fails and also add a link to query the logs of all the nodes of the chain (relaychain and the n parachains) to make mo easy to dignose what was the failure.
Also, as part of this improvements we want to build a tool that allow users to easily reproduce the job locally (using the zombienet cluster), so users don't need to figure out the images/configurations used by a failing test. The goal is to build this tool that will only receive as input the repo (e.g polkadot/cumulus/substrate/etc) and the jobId to reproduce. The tool should support both, spawn the network or run the tests from the job.
### Tasks
- [ ] Add links to grafana with error query
- [ ] Create tool to reproduce a failed test
cc @bkchr