Turns out Bamboo wasn’t working properly – I think I’d started to create a test job, and that had broken the DB patch/reboot job. Just re-ran it through and all seems good now.
Need to stop it from emailing me every night and having me just ignore the emails. I might tie it into SNS or something. It’s good to have something to play with.
Looks like my main web server instance is costing less than $2 a month on a T3 Micro. That’s a spot instance with a fairly high bidding price set so that it’s very unlikely to ever be outbid and is still the cheapest way to provision EC2 capacity.
By the same token, my database server is also a T3 micro instance and that’s looking to cost just under $7 a month on demand – same uptime as the web server so you can see there’s a $5 saving.
The bamboo server is having to run on a medium server and that’s used around $27 this month. Don’t you just love java?
Still got $300 of credit in my account so there’s no massive rush to move off the on demand instances right now. Maybe I should think about it though.
Workmail is $4 per user per month plus tax. Route53 is $0.50 per domain per month, there’s no getting around that. If I can keep my monthly bills less than $30 a month that would certainly keep my wallet happy.
Need to do a fair bit more work with bamboo to see if I can get it running on something a little less powerful or spot instance it. We’ll see.
Not a single server outage in the last three weeks. Which is great, it means that I’ve gotten everything set up properly as far as server updates, security groups and maintenance are concerned.
Bamboo seems to be doing what it’s meant to do and is rebooting the servers nightly. Everything seems to be working properly with regards to database backups and everything. I need to double check what the monthly cost is looking like when I start to run out of promotional credit at the end of this year. Could be quite costly.
Just one week until I start my new job. Excited is an understatement. Can’t wait to get going. My playing around with bamboo was in preparation for this, and it seems to be running pretty well – it’s updating and rebooting both of my ec2 instances nightly using AWS tasks and emailing me using SES to let me know whether the job has passed or failed.
Everything seems fairly stable. Bamboo itself is the biggest server I’ve got running in ec2, but it’s Java – and Java likes to eat all of the RAM so, it is what it is.
Make sure that bamboo isn’t backed off to the RDS instance you deleted, and if it is, make sure you take that final snapshot so that you can easily move it over. Oh, also be aware that mariadb isn’t compatible with bamboo unless you throw in some startup arguments.
It was a fun couple of hours at least. Yay.
I’ve started to play around with Bamboo and a few other bits and pieces in an evening and I’ve managed to find a few promotional $150 vouchers for AWS that you would usually get from AWS training courses on ebay – prices range from about £11 to £35 depending on expiry. Bamboo will not run on a T3 micro – it just doesn’t have the CPU capacity – definitely needs 2cpus, even for a tiny installation – that’s Java for you 😉
I had a spot instance running but it kept going away due to capacity. It’s running again on a spot instance but I’ve whacked up the price to $1 an hour now so hopefully that’ll never go away. But again, I’m putting database backups in place to make sure it doesn’t go anywhere, should it go down.
I’ve got $450 of credit on my account and it goes towards the EC2 cost along with quite a few other services. Here’s the complete list:
- Amazon Simple Storage Service
- AWS Lambda
- Amazon Glacier
- Amazon Sumerian
- Amazon Relational Database Service
- Amazon SimpleDB
- Amazon CloudSearch
- AWS IoT
- Amazon Lightsail
- Amazon Elastic Compute Cloud
- AWS Elastic Beanstalk
- Amazon Simple Notification Service
- AWS Config
- Amazon Redshift
- Amazon Elastic File System
- AWS Data Pipeline
- Amazon ElastiCache
- Amazon CloudFront
- Amazon Elastic Transcoder
- Amazon Virtual Private Cloud
- Amazon Lex
- AWS Data Transfer
- Amazon Simple EDI
- AWS OpsWorks
- AWS Storage Gateway
- Amazon Simple Email Service
- Amazon Machine Learning
- Amazon SageMaker
- Amazon Elastic MapReduce
- AWS X-Ray
- Amazon DynamoDB
- Amazon EC2 Container Registry (ECR)
- AWS Glue
- AWS Budgets
- Amazon Polly
- Amazon Route 53
- AWS Support (Basic)
- Amazon Elasticsearch Service
- AWS CloudTrail
- Amazon Rekognition
- AWS Key Management Service
- Amazon Simple Queue Service
- Amazon AppStream
- Amazon Kinesis
So, seeing that EC2 was in the list, and there was no way in hell I was going to burn through that much credit by the end of next year, I decided to purchase some Reserved Instances. A T3 micro for 3 years and a T3 small for one year, all upfront – total cost was $264 + 20% VAT. And this charged itself directly to my credit card, with pretty much no confirmation. To be fair to Amazon, I did click the purchase button, I just expected to see a final confirmation screen that showed me where the funds were going to be coming from. That didn’t happen.
To be fair, there is documentation to support that:
I just didn’t see it before I clicked the old purchase button 😉
Their support team have been lovely, they’ve managed to cancel the Reserved Instances and are starting the process to refund me. I upgraded to developer support at $29 a month to get my request actioned sooner. I’ve not decided whether or not I’m going to keep this in place or not yet – there are benefits to it for sure, and as I start to do more and more with it over the coming months, it might be a good idea. We’ll see.
Workmail isn’t covered by the credit, thats $4 per user per month, but to be honest, it is so much better than what I’m used to with my previous host, it makes me realise just how great an enterprise level solution it is.