# EPF5 Week-7: Data Column PeerDAS Dashboard

## Week overview
- [X] Learned about types of metrics in Grafana.
- [X] Focused on Histograms and Counters.
- [X] Dove deeper into PeerDAS specifications.
- [X] Gained an understanding of gossip and req/resp domains.
- [X] Explored the draft LH PeerDAS Grafana dashboard.
- [X] Reviewed the codebase for LH PeerDAS metrics.
- [X] Created a summary table for LH metrics.
- [X] Ran Kurtosis catching disk capacity overflow.

This week was dedicated to gaining a deep understanding of every metric. To achieve this, I delved into the gossip and request/response domains as outlined in the specifications. I explored the draft of the LH Grafana dashboard, specifically looking at how the queries for the metrics are implemented. I learned about adjusting graphs. By the end of the week, I compiled a summary table detailing the current state of LH metrics, highlighting both existing and pending metrics. Additionally, I built a data column sidecar dashboard, which I plan to discuss with the LH and DevOps teams.

### Metrics table
I compiled a summary table of all the metrics starting from Lighthouse. This table marks the current state of each metric and categorizes them, providing me with a clear overview of which metrics are ready for analysis and which ones still need attention. During this process, I discovered several metrics that were missing from the original list and identified others that should be included. I filled the table with initial values for these metrics to facilitate further analysis.


[Github issue: Add metrics to track received / sent bytes over rpc by method](https://github.com/sigp/lighthouse/issues/6114).
### Prometheus Metrics

I explored the types of Prometheus metrics: Counters, Gauges, Histograms, and Summaries. For my dashboards, I plan to use Counters, Gauges, and Histograms. I also experimented with legends, queries, time ranges, and variables. My draft Grafana dashboard currently appears as follows:

Next, I will focus on configuring settings, determining the number of nodes, and adding more metrics to it.

## Weekly EPF Activity
- [EPF5 Standup]()

- [EPF5 Office Hours]()

- [PeerDAS Breakout Room #4](https://ethereum-magicians.org/t/peerdas-breakout-4-july-23-2024/20629)

- Grandine PeerDAS Call
- [EPsg Wiki contributors meeting #5](https://github.com/eth-protocol-fellows/protocol-studies/issues/299)
Coordination of EPF wiki contributors. State of the wiki, Source Code Mirror, EPF contributions.
- [ACDC #138](https://www.youtube.com/live/lmzAUqsIbIE)
[Github issue #1100](https://github.com/ethereum/pm/issues/1100)
Devnet-1, PeerDAS, specs

## TODO
- [ ] Build a draft of LH missing network metrics.
- [ ] Begin working with Teku code.
- [ ] Adjust the existing PeerDAS metrics dashboard.
- [ ] Discuss the dashboard with my mentor devs and DevOps.

## Glossary
| Term | Description |
| -------- | -------- |
| Fork Choice | The means by which a node decides, given the information available to it, which block is the "best" head of the chain. |
| LMD Ghost | A fork choice rule used by nodes to determine the best chain. |

## Links
[EIP-7594: PeerDAS - Peer Data Availability Sampling](https://ethereum-magicians.org/t/eip-7594-peerdas-peer-data-availability-sampling/18215)
[Networking specs: gossip domain, req/resp domain](https://github.com/ethereum/consensus-specs/blob/dev/specs/phase0/p2p-interface.md)
[Upgrading Ethereum](https://eth2book.info/latest/book.pdf)
[Prometheus metrics](https://www.timescale.com/blog/four-types-prometheus-metrics-to-collect/)
[Gossip and req/resp domains](https://github.com/ethereum/consensus-specs/blob/peer-das/specs/phase0/p2p-interface.md)