nijave

44 ms·

601.
by nijave 1y ago
Iirc AWS Application Load Balancers (HTTP/L7) will cycle through IPs as fast as every 30 minutes (based on me tracking them ~5 years ago). I think they set a 10 minute ttl on their DNS records
602.
by nijave 1y ago
I think Prometheus works similar to this with some other tricks like compressing metric names. OTEL can do gRPC and a storage backend can encode that however it wants. However, I do agree it doesn't seem like efficiency was at the fore
603.
by nijave 1y ago
Yes, it allows you to bisect a program to see the block of code between log statements where the program malfunctioned. More log statements slice the code into smaller blocks meaning less places to look.
604.
by nijave 1y ago
In fairness, I think a lot of GIL languages already have high overload and I've never been under the impression OTEL was optimized for performance and efficiency.
605.
by nijave 1y ago
>but don't properly account for time spent by engineers rummaging around for data they need but don't have This is a tricky one that's come up recently. How you you quantify the value of $$$ observability platform? Anecdot
606.
by nijave 1y ago
Async replication? I think you could run semi independent regions with an orchestrator that copies config to each one. You'd go into a degraded read only state but it wouldn't be hard down. Of course bugs in the orchestrator could
607.
by nijave 1y ago
They definitely roll out code (at least for some services) one region at a time. That doesn't prevent old bugs/issues from coming up but it definitely helps prevent new ones from becoming global outages.
608.
by nijave 1y ago
A span is a discrete event emitted on completion. It contains arbitrary metadata (plus a few mandatory fields if you're following the OTEL spec). As such, it doesn't really have a beginning or end except that it has fields for dur
609.
by nijave 1y ago
You can usually turn logging on but a lot of the OTEL stack defaults to best effort and silently drops data. We had Grafana Agent running which was wrapping the reference implementation OTEL collector written in go and it was pretty easy to
610.
by nijave 1y ago
Whoops missed the autocorrect on that one facepalm
611.
by nijave 1y ago
Sure but you can still put a message up when it's some <numeric value> over some <threshold value> like errors are 50% higher than normal (maybe the SLO is 99.999% of requests are processed successfully)
612.
by nijave 1y ago
It's not rocket science. Put a message up "The service is currently degraded and some users may see errors"
613.
by nijave 1y ago
>Is it ok that my son stopped at my desk at home and saw customer PII that was left open? In practice/reality, probably. Most employers will disagree. Consider your son could just as easily over hear a phone call, see a piece of pap
614.
by nijave 1y ago
Our work SSO is set to 12/24 hours in most places which seems like a decent compromise. Auth once a day In a corporate environment, ideally your workstation password is tied to SSO and you have a short but reasonable lockscreen timeout
615.
by nijave 1y ago
Yes
616.
by nijave 1y ago
There was also an older instance with China https://www.cyberdefensemagazine.com/experts-detailed-how-ch...
617.
by nijave 1y ago
>can impact their whole infrastructure CDN and WAF seem to be working fine. I think CF rushed a lot of newer services out without the reliability some of their older/core services enjoy
618.
by nijave 1y ago
Dedupe is efficient in Borg. The target needs almost no RAM
619.
by nijave 1y ago
Isn't there some way to speed up with codegen besides using LLMs?
620.
by nijave 1y ago
This looks interesting. I've DIY'd something similar with qdirstat before but you need to arrange your emails a certain way like dated folders and can't re-slice with different criteria. On the other hand, qdirstat "cach
621.
by nijave 1y ago
You can also use an IMAP client and set it to offline/download mode so it downloads everything and saves it locally. I think "offline mode" is what it's called in Evolution--not sure what Thunderbird or other clients cal
622.
by nijave 1y ago
Why be a professional brush removal specialist when you can be a professional Fire Fighter? There's software engineers and there's enterprise software developers and companies tend to want more of the latter than the former.
623.
by nijave 1y ago
Mobile internet has underwent massive improvements since then. In the U.S. at least, there are places 5G outperforms wired connections. Same goes for mobile compute power (which translates to browser performance)
624.
by nijave 2y ago
If I'm understanding the issue correctly, it probably doesn't. From what I understand, multi-az has some setup with multiple semi synchronous replicas where only 1 replica needs to acknowledge the transaction. Aurora doesn't
625.
by nijave 2y ago
>My personal opinion is that filesystem level replication (think drbd) is the better approach for PostgreSQL That's basically what their Aurora variant does. It uses clustered/shared storage then uses traditional replication on
626.
by nijave 2y ago
It's not entirely clear but this isn't an issue in multi instance upstream Postgres clusters? Am I correct in understanding either AWS is doing something with the cluster configuration or has added some patches that introduce this
627.
by nijave 2y ago
The farther a request makes it into infrastructure, the more resources it uses.
628.
by nijave 2y ago
Or throw on TrueNAS or UnRAID if you want a GUI
629.
by nijave 2y ago
I think that'd break deleting incremental snapshots unless you tried uploading a gigantic blob of the entire filesystem, wouldn't it? Meaning you'd need to upload full snapshots on a fixed interval
630.
by nijave 2y ago
Patroni will manage a PG cluster and auto fail over. I've heard of Stolon as well. If you're on k8s, there's a couple good operators that will handle this I believe paid PG vendors like EnterpriseDB and maybe Crunchy have the

More ›