You can partition your audit/event table by time period and archive old events [1] or you can avoid the records hitting the database in the first place by generating the events elsewhere to begin with [2].
I've had a hypothesis for a while that the brain performs compacting garbage collection while it sleeps at night. While it's probably not that simple, seeing that various sibling posters are hinting at the same notion is interesting.
This type of behavior is why I can never consider GCP. How many people have been burned at this point by Google randomly shutting down something they rely on?
I've had two Google accounts shut down in the last six months with no explanation. There is no appeal. The consumer services I've used (Feed Reader, Play Music) have been shut down, and the cloud service I was most interested in was luckily shut down before I was able to use it. (They used to have a service to resize & manipulate images in Blob Storage. I found a good AWS alternative[1] instead). I cannot rely on Google for anything at all, and definitely not for something as important as cloud services.
Are there any indications to you why your accounts got shut down? Any pattern you noticed?
I - as most of us - have a personal google account, and our company uses a google business account. While I'm following news regarding google cancelling accounts at will, I fail to notice a reliable pattern: (alleged) fraud and other illegal stuff seems to comprise a good part of it, but at most 30-50%.
No, there is no pattern. The last one happened when I got a new Android phone. I logged in on my work account and my personal account, and the work account got suspended. It said "suspicious app", but the only app I used it with was Google Meet. The personal account was used for much more, but didn't get suspended. I half suspect that they deliberately have false alarms so they can act like they're more secure, but it's more likely just a horrible, unaccountable AI.
I treat all Google accounts as throwaways now and don't use the work email at all because I want to know that I can actually receive emails that are sent to me. That's a huge problem even without randomly losing access, because their spam filter has a ton of false positives and those emails don't get forwarded to my real address.
>Their spam filter has a ton of false positives and those emails don't get forwarded to my real address.
This is very interesting to me. I've used Gmail for 10 years now and I've found the spam filter to be nearly impeccable. I can't recall a single false positive. I can't even recall a single false negative, though I am moderately careful about who I provide my email to.
Now I'm left wondering if most people think about Gmail more like me or more like you...
Anyone like me who sends maybe 10 emails a month to gmail.
I'm in their spam sin bin since a spammer managed to find an old test account on my SMTP server with a weak password and spammed the world for a day or so a few years ago. The problem is that I don't send enough emails to get out of the bin.
This isn't a problem google cares about, small senders with no reputation are basically screwed. I can deliver to gmail hosted accounts I've got a relationship with (personal & my own work address) but I can't reliably send to other email addresses at work.
How do you sign up for new accounts given the mobile number requirement? Do you just reuse the same number?
I tried to make a Google account for work use the other day, and got stuck at that point. Given Google's history it seems silly to use my personal account for work, or to connect the two accounts in any way.
I had two personal accounts that I registered before GMail had a phone requirement. I don't remember setting up the work account, but I probably just used my normal phone number.
Play Music has not been shut down (yet), and you can transfer everything to Youtube Music, which is available at the same price (and in my opinon a superior product).
Some randomly selected people can transfer everything to YouTube Music. I can't, and it may be months before Google would allow me to. It's exactly that kind of treatment that makes me feel like Google has zero respect for its customers.
Spotify is generally better than Play Music though, so it was for the best in the end.
a) Spy on people and sell the data to advertisers.
b) Use that data to directly push ads
That's basically incompatible with b2b services. Or consumer services. As a customer you're judged by how valuable the data they are collecting on you is. Which is less than a support call costs. That bleeds into every facet of their business. As such even if you pay them money you get the same treatment because they can't think any different.
Advertisers can choose to advertise to customers that fall into various buckets.
Income, gender, location, recent history of viewing specific types of pages, etc.
This is what I mean by limited form of access. Advertisers do not receive user information, but are granted the ability to use it when setting up ad campaigns.
They don't sell peoples personal data at least publicly. But that makes them care even less about end users, even paying end users. Which means a business is foolish to rely on them.
I know Ethereum uses Merkel Trees to store state but this app already has Postgres and I imagine that it would be better architecturally to not write a big blob data structure back and forth. WITH RECURSIVE queries and the right table layout could make this entire thing way simpler to operate.
Feels as though this essay was written by Ned Ludd himself. Software has been constantly commodified and yet the hurdle the market expects you to clear gets higher at the same time.
I whole-heartedly agree and this is why I give money to AWS and Azure will not give any to GCP until the lack of transparency and random product killings stop.
We were quite unhappy with Aurora because of the terrible performance of writes against tables with secondary indices. What does your workload look like?
We trimmed out most of our secondary indexes long before we moved to Aurora because performance of postgres writes with secondary indexes is always a bit rough. The few we have remaining do well enough.
Overall our workload is very read heavy. At peak, if we compare our CPU on the writer vs the aggregate CPU on the readers, we do about 10x more read work than writes.
Remind is an education messaging program, so our workload is partially user management (which users belong to which schools and which classes) and partially user generated content (messages being sent). Our user generated content (more like 2-3x read vs write) is all backed by DynamoDB and our user management is in a couple of Aurora database clusters.
Too late now, but would it have been useful/possible to go public about it or threaten to talk the media and cause way worse problems for them? Considering that security is their game, I wonder if the bad PR would've made them re-think targeting you.
[1] https://github.com/icheishvili/audit-trigger [2] https://github.com/eulerto/wal2json