Andreas K@_elena Well, AFAIK, ddg uses bing as their primary index.
Actually, the incestuous relationship with Microsoft has been commented on, e.g. ddg protects against anything BUT Microsoft.
Not really news.
@_elena And Signal being down because of AWS?
a) who is surprised, where do you think do they host their stuff? On their private cable modem?
b) but seriously, Signal might be down even indirectly, e.g. they use a service (that they know is critical, or that just proved to be more critical than they thought), which uses AWS.
I remember a decade ago, when the SOLR servers went on strike. Everybody (we were a small company at the time), dropped everything and started trying to figure that one.
different kind of backups, VM snapshots etc all did not start up. Everybody tearing their hairs out.
Hours later: Oh, our logging service refused service (I don't remember actually if it was because the account went into overage, whatever). But because the way the stuff was configured, SOLR refused to startup and to give extra spice to it, the messages it spit out were not exactly helpful.
Point is, just because X fails when Y is down, X does not need to depend upon Y directly. In our complex
world nowadays, it's quite possible that Y is down for completely not related reasons. Or the via an indirect dependency pathway that nobody thought about.
(That can be stuff as stupid that something being down shifts traffic pattern and thus acts as an unintended DDoS on something your system requires.)
(That can be stuff as stupid that something being down shifts traffic pattern and thus acts as an unintended DDoS on something your system requires.)
Roy@yacc143
Excellent points about non-obvious dependencies that can be easily overlooked during implementation.
Thanks for the reminder.
Excellent points about non-obvious dependencies that can be easily overlooked during implementation.
Thanks for the reminder.
@roytoo Nothing that teaches one that little gem as spending a day on adrenaline trying to resolve an issue, and the next day during the standup the DevOps guy quips that it was the logging.
But the uglier part is that in our “service” oriented world, you often only have a contract (legal one) that describes what your service provider will do or not. And that is under the naive assumption that the service provider will mostly honour the contract.
Now a corporation is not a single entity,
and it's totally and absolutely possible that you buy some service, they even promise (legally) that they won't use AWS, and only 1 commercial acquisition of your service provider, + 12-18 months (you have to factor in the times to “merge the infrastructure” to enjoy the synergies of the acquisition), and your service provider provides the service via the infrastructure of its new parent company, which is a wholly owned subsidiary of Amazon.
Happens even at a "consumer" level. I remember choosing around 2 decades ago a smaller, slightly more expensive ISP, well, to avoid the local cable monopoly. 2-3 years later on, they were bought by the cable company. Bad enough, but one learns to live with it.
A couple of years ago, our cable company (who would have thought that I would remember them in such positive ways), was acquired by T-Mobile. And yet another run around the enshitification spiral.
As explanation for non-locals (e.g. Americans), T-Mobile pretends to be a “premium” brand in our market here, you know we are not only a virtual operator, we have our real own infrastructure, and look how great it is.
But their user service reached the point, that generally discount brands have better service.