hckrnws

I do not work in the space at all, but it seems like Cloudflare has been having more network disruptions lately than they used to. To anyone who deals with this sort of thing, is that just recency bias?

Launching a new service every 5 minutes is obviously stretching their resources.

It is not. They went about 5 years without one of these, and had a handful over the last 6 months. They're really going to need to figure out what's going wrong and clean up shop.

Engineers have been vibe coding a lot recently...

The featured blog post where one of their engineers presented an allegedly "production grade" Matrix implementation in which authentication was stubbed out as a TODO says it all really.

Thats a classic claude move, even the new sonnet 4.6 still does this.

It has been roughly speaking five and a half years since the IPO. The original CTO (John Graham-Cumming) left about a year ago.

They coasted on momentum for half a year. I don't even think it says anything negative about the current CTO, but more of what an exception JGC is relative to what is normal. A CTO leaving would never show up the next day in the stats, the position is strategic after all. But you'd expect to see the effect after a while, 6 months is longer than I would have expected, but short enough that cause and effect are undeniable.

Even so, it is a strong reminder not to rely on any one vendor for critical stuff, in case that wasn't clear enough yet.

I wondered what happened to him?

Cloudflare Outages are as predictable, as the Sun coming up tomorrow. Its their engineering culture.

https://hn.algolia.com/?dateRange=all&page=0&prefix=false&qu...

Insufficient mock data in the staging environment? Like no BYOIP prefixes at all? Since even one prefix should have shown that it would be deleted by that subtask...

From all the recent outages, it sounds like Cloudflare is barely tested at all. Maybe they have lots of unit tests etc, but they do not seem to test their whole system... I get that their whole setup is vast, but even testing that subtask manually would have surfaced the bug

I think Cloudflare does not sufficiently test lesser-used options. I lurk in the R2 Discord and a lot of users seem to have problems with custom domains.

It was also merged 15 days prior to production release...however, you're spot on with the empty test. That's a basic scenario that if it returned all...is like oh no.

While neither am I nor the company I work for directly impacted by this outage, I wonder how long can Cloudflare take these hits and keep apologizing for it. Truly appreciate them being transparent about it, but businesses care more about SLAs and uptime than the incident report.

I’ll take clarity and actual RCAs than Microsoft’s approach of not notifying customers and keeping their status page green until enough people notice.

One thing I do appreciate about cloudflare is their actual use of their status page. That’s not to say these outages are okay. They aren’t. However I’m pretty confident in saying that a lot of providers would have a big paper trail of outages if they were more honest to the same degree or more so than cloudflare. At least from what I’ve noticed, especially this year.

Azure straight up refuses to show me if there's even an incident even if I can literally not access shit.

But last few months has been quite rough for Cloudflare, and a few outages on their Workers platform that didn't quite make the headlines too. Can't wait for Code Orange to get to production.

Bluntly: they expended that credit a while ago. Those that can will move on. Those that can't have a real problem.

As for your last sentence:

Businesses really do care about the incident reports because they give good insight into whether they can trust the company going forward. Full transparency and a clear path to non-repetition due to process or software changes are called for. You be the judge of whether or not you think that standard has been met.

is this blog post LLM generated?

the explanation makes no sense:

> Because the client is passing pending_delete with no value, the result of Query().Get(“pending_delete”) here will be an empty string (“”), so the API server interprets this as a request for all BYOIP prefixes instead of just those prefixes that were supposed to be removed. The system interpreted this as all returned prefixes being queued for deletion.

client:

     resp, err := d.doRequest(ctx, http.MethodGet, `/v1/prefixes?pending_delete`, nil)

    if v := req.URL.Query().Get("pending_delete"); v != "" {
        // ignore other behavior and fetch pending objects from the ip_prefixes_deleted table
        prefixes, err := c.RO().IPPrefixes().FetchPrefixesPendingDeletion(ctx)
        if err != nil {
            api.RenderError(ctx, w, ErrInternalError)
            return
        }

        api.Render(ctx, w, http.StatusOK, renderIPPrefixAPIResponse(prefixes, nil))
        return
    }

> even if the client had passed a value it would have still done exactly the same thing, as the value of "v" (or anything from the request) is not used in that block

If they passed in any value, they would have entered the block and returned early with the results of FetchPrefixesPendingDeletion.

From the post:

> this was implemented as part of a regularly running sub-task that checks for BYOIP prefixes that should be removed, and then removes them.

They expected to drop into the block of code above, but since they didn't, they returned all routes.

yep, no mention that re-advertised prefixes would be withdrawn again as well during the entire impact even after they shut it down.

doesn't look AI-generated. even if they have made a mistake, it's probably just from the rush of getting a postmortem out prior to root cause analysis

The eternal tech outage aphorism: It's always DNS, except for when it's BGP.

This blog post is inaccurate, the prefixes were being revoked over and over - to keep your prefixes advertised you had to have a script that would readd them or else it would be withdrawn again. The way they seemed to word it is really dishonest.

The irony is that the outage was caused by a change from the "Code Orange: Fail Small initiative".

They definitely failed big this time.

Another SPoF found, experimentally. If these things reproduce, the problem is probably deeper than one particular bug or typo.

https://replicated.wiki/blog/args.html

DaaS - Downtime as a Service©

Just joking, no offence :)