> The OVHcloud control panel is a labyrinth: the lifecycle rule configuration is buried somewhere in the documentation, and it involves some work in the terminal.
Use OpenTofu/Terraform! Much better than messing with cloud consoles, and then your infrastructure self-documents.
I’d also put out one note to any people outside the EU looking to switch to Mistral or really any service: just because they’re a European company doesn’t mean they’ll follow the GDPR if you don’t live there. Mistral is an example: in their privacy policy, they state that they follow whatever privacy laws exist in your country.
> they state that they follow whatever privacy laws exist in your country
Well, that's kinda obvious - if they want to do business in a country, they have to follow the laws of that country. That doesn't in and of itself mean that they will apply weaker privacy protections if the local laws are less strict than GDPR...
Matomo is nice on low traffic, but when we have a sustained rate of 5-25 logins per second and above things become real slow. Using regexps is really bad when you start having problems, but they are fine on low traffic sites.
So If like it but it is a headache on high traffic sites. If anyone have an easy solution I would gladly accept it.
I have worked with two clients. Both north of 8 million visits a month. Both on matomo. Both self hosted.
If you architect the underlying infra right it still works like a charm. But I admit people need to know what they are doing. I was quite impressed with both infra teams.
But as always, if you do not want tu use auto scaling US cloud based services, you need to enasure you have the right scaling and the necessary technical expertise at hand.
Managed a fairly large matomo site in the past. Using queue plugin (https://plugins.matomo.org/QueuedTracking) with Redis Cluster really improves the situation. We actually built a custom plugin with Nginx + Lua to avoid PHP altogether for the tracking part. Scaling ingestion then wasn't the problem, draining the queue was
> Digital sovereignty sounds like a buzzword until you think
Sure now just think and give me the reason. All these moving to Europe post is getting tiring. Amazon follows the same EU rules, if not more, than Scaleway.
Matt Lakeman writes in one of his blogs that wherever he goes, people tend to love the USA. Except in Europe where he faces a constant storm of criticism. And that was before February. Just like you cannot explain the taste of chocolate to someone, it is hardly possible to explain the mental shift that happened everywhere when the US threatened the EU with military invasion. Like a broken egg this is diplomatic damage that cannot be repaired.
If you sell software and you tell your customers and prospects that everything runs in Europe, by European companies, this instills an enormous amount of trust. Risk averse sectors like manufacturing love this, and it will help you gain customers immediately.
So no, these posts are not tiring to many of us. In fact, we are only at the beginning of the beginning because many of us will be making these migrations. I wish things had run a different course.
They had a datacenter burn down (in large part because it was fully built using wood) and lost all customer data and did not take any action for 6 months after the incident.
I've been dabbling with OVH and it feels very pricey and fragile. Has a very lipstick on a pig approach to whatever they used to be doing before piling into cloud.
While the incident did happen, a lot of actions were taken and most of the data was recovered.
OVH now also keeps backups even for clients that don't pay for it.
I was hit by that datacenter catastrophe and got my data back almost immediately, in a new VM.
I've been using them for years with little issue (no more than happened on my AWS or Azure accounts, I would say less because it's less of a mess in general).
It's fine to have an unstable backup system, as long as any failures in your backups are uncorrelated with failures in your primary system. And a random datacentre burning down probably isn't correlated with anything else, unless you're foolish enough to host your primary and backup copies in the same building.
All else equal, a more stable backup is of course better, but any backup is better than no backups, so choosing the cheapest possible option is often the best strategy since that's the one that you're the most likely to keep using long-term.
That lettermint service looks interesting! I was recently looking for something in that price range that covers both transactional and broadcast emails but couldn't find anything in Europe so I settled on Postmark which has been good, this looks almost identical in features and pricing though.
Same here. I just discovered this and put it in my "check out tonight" folder. I am currently happy using resend. But this looks interesting, especially also for my freelance clients with a focus on EU tech.
Yes, Lettermint looks great if you need both transactional and marketing emails. But it's a bit weird that the OP selected Lettermint over Scaleway for just transactional emails when they use lots of other Scaleway services.
While I agree with him that the US is becoming more unpredictable, I don't think the EU is much better, especially with regards to digital things where they can be worse in some ways. For example, they are discussing restricting VPN access for 'child protection'[1]
Matomo charges 22 euros for 50k hits/month.[0] Basically, it's unusable for anything other than a hobby site - especially with the number of crawlers nowadays.
If you self host for free, you're missing basically all of the good parts of web analytics such as funnel analysis as they lock all of those features being paid subs.
Why are there exceptions for Anthropic, GitHub and GitLab?
> Anthropic is a US company...But it satisfies something else, the sense that the organization building the thing has given serious thought to what it’s building and why.
This reads like a weak excuse. Mistral and Mistral Vibe exists and even if you don't like them, there are many non-US harnesses (Qwen code) that are available.
> GitHub stays in the picture for one specific purpose: public-facing NPM packages and issue tracking for open source software.
First of all Codeberg exists.
Secondly, at this stage relying on NPM and the Java/Typescript ecosystem is quite frankly waiting for a disaster to happen.
This post isn't absolute on moving their digital stack to Europe as it has not one but three exceptions too many.
For the past days I've been participating(albeit over Teams) in a conference relevant to my industry (intel), basically startups and established companies showcasing their products to a closed audience of EU gov. officials.
One thing I noticed right away, is that all companies were asked "Can we fully host this from within EU or our country" from the various people in audience. Every single one. Many of the startups had slides prepared for this.
Definitely a change, because it is not something I can recall being important just a couple of years ago.
I’ve been working in a European SaaS for a few years and I’ve even noticed we get more queries about using specially EU providers. In the past AWS eu-west was sufficiently “In Europe” for clients that cared about that but there’s a growing minority that would like us not to use AWS at all. Not in large enough numbers that our company is planning a migration but it’s being discussed in a way it wasn’t in 2024.
Just a nitpick: 1Password is Canadian (still not European, but not us based, if that’s the issue). I do understand the choice to move all into proton though.
I switched to Protonmail a month ago. It is patently inferior to gmail. Every day I get annoyed by some weakness in the UI that google had apparently just always solved without me ever having to think about it. For example, reading long email chains in the proton UI is horrific. I don't know what google did that made it natural to read and proton does so badly, but it is painful to read these long chains of emails. Another example is log emails from my servers are getting grouped together by Proton. Gmail had sepearated the logs into separate emails in a very natural way. These small annoyances add up and I'm not having a fun time right now with proton.
A pragmatic article, always nice. I was surprised that gitlab and github was stillton the list. For me moving to self hosted forgejo was one of the easiest transition i had. But i did not have complex CI/CD needs
While I do think the link highlights are pretty neat, this particular cursor hijack annoys me greatly. Would be nicer to float the link highlights next to the standard cursor.
Proton Mail not supporting filters for message bodies is brutal, I understand why they don't do it but that really lowers its usability for me. Bummer.
144 comments
[ 2.9 ms ] story [ 96.8 ms ] threadUse OpenTofu/Terraform! Much better than messing with cloud consoles, and then your infrastructure self-documents.
I’d also put out one note to any people outside the EU looking to switch to Mistral or really any service: just because they’re a European company doesn’t mean they’ll follow the GDPR if you don’t live there. Mistral is an example: in their privacy policy, they state that they follow whatever privacy laws exist in your country.
Well, that's kinda obvious - if they want to do business in a country, they have to follow the laws of that country. That doesn't in and of itself mean that they will apply weaker privacy protections if the local laws are less strict than GDPR...
So If like it but it is a headache on high traffic sites. If anyone have an easy solution I would gladly accept it.
If you architect the underlying infra right it still works like a charm. But I admit people need to know what they are doing. I was quite impressed with both infra teams.
But as always, if you do not want tu use auto scaling US cloud based services, you need to enasure you have the right scaling and the necessary technical expertise at hand.
Sure now just think and give me the reason. All these moving to Europe post is getting tiring. Amazon follows the same EU rules, if not more, than Scaleway.
more mean the US rules that hoover up all the data for the government
If you sell software and you tell your customers and prospects that everything runs in Europe, by European companies, this instills an enormous amount of trust. Risk averse sectors like manufacturing love this, and it will help you gain customers immediately.
So no, these posts are not tiring to many of us. In fact, we are only at the beginning of the beginning because many of us will be making these migrations. I wish things had run a different course.
They had a datacenter burn down (in large part because it was fully built using wood) and lost all customer data and did not take any action for 6 months after the incident.
They're just not a serious company.
While the incident did happen, a lot of actions were taken and most of the data was recovered. OVH now also keeps backups even for clients that don't pay for it.
I was hit by that datacenter catastrophe and got my data back almost immediately, in a new VM.
I've been using them for years with little issue (no more than happened on my AWS or Azure accounts, I would say less because it's less of a mess in general).
Stop spreading false rumors.
All else equal, a more stable backup is of course better, but any backup is better than no backups, so choosing the cheapest possible option is often the best strategy since that's the one that you're the most likely to keep using long-term.
https://www.scaleway.com/en/transactional-email-tem/
[EDIT] Found the answer from the OP:
> To be honest I implemented Lettermint before I migrated to Scaleway, so I didn't even look at TEM
[1] https://www.europarl.europa.eu/thinktank/en/document/EPRS_AT...
Matomo charges 22 euros for 50k hits/month.[0] Basically, it's unusable for anything other than a hobby site - especially with the number of crawlers nowadays.
If you self host for free, you're missing basically all of the good parts of web analytics such as funnel analysis as they lock all of those features being paid subs.
[0]https://matomo.org/pricing/
Why are there exceptions for Anthropic, GitHub and GitLab?
> Anthropic is a US company...But it satisfies something else, the sense that the organization building the thing has given serious thought to what it’s building and why.
This reads like a weak excuse. Mistral and Mistral Vibe exists and even if you don't like them, there are many non-US harnesses (Qwen code) that are available.
> GitHub stays in the picture for one specific purpose: public-facing NPM packages and issue tracking for open source software.
First of all Codeberg exists.
Secondly, at this stage relying on NPM and the Java/Typescript ecosystem is quite frankly waiting for a disaster to happen.
This post isn't absolute on moving their digital stack to Europe as it has not one but three exceptions too many.
One thing I noticed right away, is that all companies were asked "Can we fully host this from within EU or our country" from the various people in audience. Every single one. Many of the startups had slides prepared for this.
Definitely a change, because it is not something I can recall being important just a couple of years ago.
Off topic: that’s a beautiful website
No ddos protection yet.