Review client type distribution collection

[philknows]

Expected Behavior

Currently the crawler is pulling a count for the Client type distribution metric which unfortunately does not look accurate in distribution. Other crawlers such as clientdiversity.org is reporting very different distributions using Miga Labs Armiarma and Sigma Prime's Blockprint crawlers.

Might be related to #205.

Current Behavior

Numbers look inaccurate.

Possible Solution

We should be clear in how we are capturing this metric. It may be good to outline here how this metric is captured and create some sort of tooltip in the UI to explain what is being done to get this number.

[spacesailor24]

From my understanding, we're asking each node to report it's AgentName (code here) like Miga Labs' crawler, so our numbers should probably be closer to theirs. I think it would make sense that the lack of a digest filter (#205) would be causing our skewed numbers. I think the best way to test this would be to connect an API with the fork digest filter to our prod DB and apply corresponding mainnet fork digests and see if the numbers are closer to what Miga Labs is reporting

┌─────────┬──────────────┬──────────────┬──────────────┐
│ (index) │    phase0    │    altair    │  bellatrix   │
├─────────┼──────────────┼──────────────┼──────────────┤
│ mainnet │ '0xb5303f2a' │ '0xafcaaba0' │ '0x4a26c58b' │
│ goerli  │ '0x79df0428' │ '0x82f4a72b' │ '0xc2ce3aa8' │
│ ropsten │ '0x4be89531' │ '0xf56931b6' │ '0x3cfa3bac' │
│ sepolia │ '0xa8fee8ee' │ '0x60725ca1' │ '0x36fa5013' │
└─────────┴──────────────┴──────────────┴──────────────┘

[sadiq1971]

Right. A test with the updated api on prod will give us more accurate result. I am also waiting for testing.