Add Retries to ENSRainbow fetches

[shrugs]

TypeError: fetch failed
    at labelByLabelHash (/app/apps/ensindexer/src/lib/graphnode-helpers.ts:28:24)
    at /app/apps/ensindexer/src/handlers/Registry.ts:188:23
  26 |  **/
  27 | export async function labelByLabelHash(labelHash: LabelHash): Promise<Label | null> {
> 28 |   const healResponse = await ensRainbowApiClient.heal(labelHash);
     |                        ^
  29 |
  30 |   if (!isHealError(healResponse)) {
  31 |     // original label found for the labelHash
caused by: Error: read ECONNRESET
    at TCP.onStreamRead (node:internal/stream_base_commons:216:20)

ENSRainbow client should gracefully retry (with backoff) over the course of like 5-10 seconds before dying

[lightwalker-eth]

@shrugs Thanks for flagging this. We were working on this goal in this PR: #344 It just got a bit overlooked as we were pushing so aggressively on other goals. For example, see the isRetryableHealError that was created in the PR.

How urgent do you see this issue?

[shrugs]

the next release will likely be just the update of ponder to fix crash recovery, but it should be in the next release, we've hit this a few times, and since ensrainbow will always restart deploy after ensindexer starts during a deploy, we hit this codepath every deploy so might as well handle it correctly

[shrugs]

closing in favor of #344