hakon.eth

@hakon

54 Following

79 Followers

hakon.eth pfp

Ahh fixed. And thanks I guess 🥲

1 reply

0 recast

1 reaction

hakon.eth pfp

I no longer work in Dune so I can’t help unfortunately 😔

1 reply

0 recast

1 reaction

hakon.eth pfp

@mvr how does the data look like now? I think I fixed it. Was unfortunately much much harder than I thought.

1 reply

0 recast

1 reaction

hakon.eth pfp

Hey sorry for the late response, been busy touching grass all weekend (just became the father of two). We saw some schema inconsistentcies with the data on Friday which we couldn't resolve in time. Overall I hope you are happy with the new and more frequent updates, except for this hickup

2 replies

0 recast

1 reaction

hakon.eth pfp

Thanks to my awesome team!

0 reply

0 recast

1 reaction

hakon.eth pfp

Right, this should be fixed now, can you confirm?

1 reply

0 recast

1 reaction

hakon.eth pfp

A bit unsure. We'll have to ask Neynar. New sync is getting close to production and that will be incremental for sure.

0 reply

0 recast

3 reactions

hakon.eth pfp

1 reply

0 recast

2 reactions

hakon.eth pfp

I'll just write a bot to reply! We're seeing the same issues as before but they are more pronounced every day. We run syncs and they fail after many many hours and we have to start all over. We started working on a much better and frequent solution on Tuesday so stay tuned!

4 replies

0 recast

4 reactions

hakon.eth pfp

Hi! Can you elaborate on what is problematic, inaccurate, or buggy? As far as I know the latest snapshot should be complete and be from yesterday 8PM UTC.

0 reply

0 recast

0 reaction

hakon.eth pfp

I see green lights in our infra again, and I believe the datasets should be up to date.

1 reply

0 recast

1 reaction

hakon.eth pfp

I'd be careful pointing fingers at Neynar in this case. Yes, the data delivery is out of our control, but it is not because they provide poor service. We chose a quick-win integration with a full sync from a database rather than doing an incremental sync which would have taken longer.

1 reply

0 recast

2 reactions

hakon.eth pfp

Not good unfortunately. Everything should be fixed on our end, but we've had timeouts from the Neynar data source, out of our control. We're going to discuss how to fix this once and for all this week.

1 reply

0 recast

0 reaction

hakon.eth pfp

I renamed them now. Apparently we had a latent bug/config drift.

1 reply

0 recast

1 reaction

hakon.eth pfp

0 reply

0 recast

0 reaction

hakon.eth pfp

Old results? I get this.

2 replies

0 recast

0 reaction

hakon.eth pfp

@mvr I saw this in our infra this morning.

3 replies

0 recast

0 reaction

hakon.eth pfp

0 reply

0 recast

2 reactions

hakon.eth pfp

@mvr we just deployed a fix. The datasets sometimes time out though and there is nothing we can do short term. The whole thing starts to get a bit embarrassing but to be honest the alternative at the time would be to not sync the dataset at all. We're trying to be mean and obviously there's demand for this dataset.

3 replies

0 recast

2 reactions

hakon.eth pfp

Okay, updated the pointers to the latest data we have. Which unfortunately is not super up to date, we had some crashes last night.

4 replies

0 recast

1 reaction