# EPF5 Week-16: Draft Summary Dashboard

## Week overview
- [X] Updated the dashboard for Kurtosis.
- [X] Built a draft summary dashboard based on the metrics specs, created filters for it.
- [X] Proceeded with Lighthouse PRs.
- [X] My [Grandine](https://github.com/hangleang/grandine/pull/2) and [Teku](https://github.com/Nashatyrev/teku/pull/139) PRs have been merged.
- [X] Got stuck with the Teku histogram implementation.

This week, I kept working on metrics for different clients. Some of my PRs were merged. I also made a draft summary dashboard for all clients on DevOps Grafana.

### Grafana
I built a draft dashboard on DevOps Grafana. It shows a summary of the metrics have been added to CL clients. Right now, it has some PeerDAS metrics for Lighthouse, Grandine, and Teku. I added a filter so each client can see just their own numbers, or compare with others.

### Lighthouse
My Lighthouse [Req/Resp](https://github.com/sigp/lighthouse/pull/6431) and [Gossipsub](https://github.com/sigp/lighthouse/pull/6430) PRs didn't pass CI tests at first. So I fixed the code. Now all tests work except the Ubuntu test. I'm waiting for someone from the team to review it and hopefully merge it soon.

### Grandine
My PR with the first batch of metrics for Grandine got merged. On our last Grandine PeerDAS call, we decided to test these locally using Kurtosis, then on peerdas-devnet-2.

### Teku
This week, I had trouble adding histogram metrics to Teku. Unlike Lighthouse and Grandine, Teku doesn't usually use histograms. I found one example where they used [HDR histogram](https://github.com/HdrHistogram/HdrHistogram), but it won't work for me. So I looked into different metric libraries for Java. The recommended library is this [Prometheus Client Java Library](https://github.com/prometheus/client_java). So, I am going to proceed with it. I'll keep working on the histogram metrics implementation because half of the beacon PeerDAS metrics need histograms. It's important to add this to Teku.

As a side job, I added a CLI [floodPublish flag](https://github.com/Consensys/teku/pull/8610) to Teku. It will be in the next release. This should help homestakers who don't have much bandwidth. Enrico wrote about it in detail. The details are in [Enrico's post](https://hackmd.io/@dsjdqQDBRp66btjRzfxjbw/r1OuFG7AA).

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

- [EPF5 Office Hours](https://github.com/eth-protocol-fellows/cohort-five/issues/452)

- [ACDE #197](https://youtube.com/live/PWhn8KdgCl8)
[Github issue #1153](https://github.com/ethereum/pm/issues/1153)
devnet-3, EIPs 7702, 7762, Pectra forksplit discussions

## TODO
- [ ] Continue implementing histogram metrics on Teku
- [ ] Create a dashboard for Grandine
- [ ] Attend PeerDAS Breakout Room to discuss metrics specs

## Glossary
| Term | Description |
| -------- | -------- |
| Native Histogram | A native histogram requires only one time series, which includes a dynamic number of buckets in addition to the sum and count of observations. Native histograms allow much higher resolution at a fraction of the cost. |

## Links
[Prometheus Histogram](https://prometheus.io/docs/concepts/metric_types/#histogram)
[Prometheus Client Java Library](https://github.com/prometheus/client_java)
[Flood Publishing](https://github.com/libp2p/specs/blob/master/pubsub/gossipsub/gossipsub-v1.1.md#flood-publishing)