kondro

14 ms·

121.
by kondro 4y ago
It just seems to be a management platform to allow you to deploy and track the runtimes of an unlimited number of Workers (i.e. functions) that operate in a tenant-isolated manner. This is surprisingly easy to do with AWS Lambda right now,
122.
by kondro 4y ago
Great. :)
123.
by kondro 4y ago
Curious about the costs of this. Wouldn't it cost at least $13/month just in PutObject request costs to replicate Sqlite to S3 at the default of 1 sync per second? Or is it smart enough to only sync if there have been additions to
124.
by kondro 4y ago
Sorry if I wasn’t clear, but I meant specially the CDN product.
125.
by kondro 4y ago
Except this feels different from what was promised at launch, before the radio silence for the last 9 months. Except I can't get unlimited egress, because the product isn't available yet. Egress isn't our price pain point any
126.
by kondro 4y ago
Another Cloudflare announcement without releasing the actual product. Why can't they wait until they have something to actually show rather than announcing something that we can't have. R2 was announced in September 2021 and still
127.
by kondro 4y ago
After trying everything and always going back to the native Gmail experience I recently switched to https://mimestream.com/ and have been loving it.
128.
by kondro 5y ago
In Ruby, but https://bullettrain.co/ is really awesome. Billing, teams, roles, chat, outbound webhooks, etc
129.
by kondro 5y ago
Wasabi's policy prevents reading more than you write. They're designed for backup, not publication.
130.
by kondro 5y ago
Some of the fastest memory available for the entire system with zero-copy to the GPU. A lot of what a discrete GPU needs to do is move bits backwards and forwards with the main CPU/RAM.
131.
by kondro 5y ago
Just curious, how are indexes priced? I assume you’ll pay $2.50/GB for storage, but if I update a row that touches 5 indexes is this 1 write or 6?
132.
by kondro 5y ago
The minimum size of objects in cheaper storage types is 128KiB. Given the article quotes $100k to run an inventory (and $100k/month in standard storage) it's likely most of your objects are smaller than 128KiB and so probably woul
133.
by kondro 5y ago
More like 15,000 if you average 1TB per customer per month with 50%-ish burst.
134.
by kondro 5y ago
You don't need to buy the lossless plan.
135.
by kondro 5y ago
JetPens have a number of sampler packs: https://www.jetpens.com/search?q=sample&v=2
136.
by kondro 5y ago
But what did you actually assign? EBS volumes are priced based on provisioned capacity (including throughput if you choose io2 volumes).
137.
by kondro 5y ago
These prices all feel like standard MacBook Pro pricing to me. I just ordered one (M1 Max, 64GB, 2TB) and it cost as much as that tier always seems to cost me.
138.
by kondro 5y ago
At least you can now tell CFN not to rollback on failure… definitely helps when developing your CFN.
139.
by kondro 5y ago
Great pricing, but a 48 hour retrieval window sucks for everything except things you basically never need to retrieve. I'm not even talking backups here, if you need a backup at all, you probably need it ASAP. This is a governance prod
140.
by kondro 5y ago
All the alternatives for 11 9's durability object stores have similar prices to S3. There aren't any alternatives. Wasabi gets close, but their 1:1 storage to egress ratio makes it great for backups and a small subset of use-cases
141.
by kondro 5y ago
Don't forget AWS CloudWatch for logs. Cost for log processing is $0.50 per GB and the minimum size of logs (just for the START/END/REPORT lines output by Lambda itself, before you start logging any of your data) is about 26
142.
by kondro 5y ago
An easy to digest article: https://authzed.com/blog/what-is-zanzibar/ tl;dr: Highly scalable RBAC/ABAC
143.
by kondro 5y ago
They were first and from a featureset and reliability perspective, S3 is still unparalleled. That coupled with storage costs that were always very competitive and the fact you had unlimited scale and PAYG pricing got a lot of people hooked.
144.
by kondro 5y ago
No. The all-inclusive Lambda workers are limited to 50ms of actual CPU runtime and can execute forever (i.e. hours) for IO bound workloads, as long as you stay below the 50 network requests per execution. And for that they cost $0.50&#x2F
145.
by kondro 5y ago
When AWS was a baby charging around $0.10/GB for egress I had a quote sitting on my desk from Akamai wanting $1.50/GB for a minimum of 1TB per month. They were (are?) even more addicted to their egress charges.
146.
by kondro 5y ago
Which are often significantly lower than request (and egress) charges. 1,000 reads per second is 2.6 billion per month. At S3 prices, that wound cost around $1,040. 100x storage is pretty minimal to achieve this. For example: 100x 128KiB =
147.
by kondro 5y ago
I don’t think they mentioned this would be “per object” and I assumed they meant per account. Free per object requests is super game-able. Want 1000 free requests per second, just rotate through 100 identical objects.
148.
by kondro 5y ago
Pretty similar really. Latency is much less important than throughput for all of my use-cases.
149.
by kondro 5y ago
It wasn't the code that was the issue (and you probably wouldn't want an S3 compatible API). It's that on top of the Durable Store pricing, you need to pay for the $0.15/million requests + $12.50/million GB-sec proc
150.
by kondro 5y ago
Me too… but I don't think anything really does this. _Maybe_ Firebase Store for certain scenarios, but it has its own tradeoffs too.

More ›