CloudFront caching strategies
Cache keys and invalidation. This post digs into the topic, drawing on recent discussion from Hacker News, Reddit, and the wider community.
The conversation that prompted this post is linked in the sources. I have tried to capture the substance of the discussion rather than just summarize it, because the comments are usually where the real insight lives.
What happened
Autoscaling is not a substitute for capacity planning. It reacts to load; it does not predict it.
Observability is the difference between knowing something is broken and knowing why. Invest in it before you need it.
Backups are not disaster recovery. Test your restores or you do not have backups.
Tagging is boring and essential. Without consistent tags, cost attribution is guesswork.
The cloud is just someone else's computer, and that someone else is very good at charging you for convenience. The question is always whether the convenience is worth the markup.
Reserved instances and savings plans reward commitment. If your workload is stable, not using them is leaving money on the table.
Infrastructure as code is not optional at scale. Clicking in the console is fine for learning and unacceptable for production.
Networking is where cloud bills hide. Data transfer charges between regions and out to the internet add up fast.
The cloud is just someone else's computer, and that someone else is very good at charging you for convenience. The question is always whether the convenience is worth the markup.
Serverless is great until it isn't. Cold starts, concurrency limits, and vendor lock-in are real costs that show up at scale.
Serverless is great until it isn't. Cold starts, concurrency limits, and vendor lock-in are real costs that show up at scale.
Networking is where cloud bills hide. Data transfer charges between regions and out to the internet add up fast.
Spot instances are free money if your workload can tolerate interruption. Most batch and async workloads can.
Infrastructure as code is not optional at scale. Clicking in the console is fine for learning and unacceptable for production.
Reserved instances and savings plans reward commitment. If your workload is stable, not using them is leaving money on the table.
Why it matters
The shared responsibility model means the cloud secures the infrastructure and you secure everything else. Most breaches are in the everything else.
Backups are not disaster recovery. Test your restores or you do not have backups.
Autoscaling is not a substitute for capacity planning. It reacts to load; it does not predict it.
Spot instances are free money if your workload can tolerate interruption. Most batch and async workloads can.
Networking is where cloud bills hide. Data transfer charges between regions and out to the internet add up fast.
Tagging is boring and essential. Without consistent tags, cost attribution is guesswork.
Reserved instances and savings plans reward commitment. If your workload is stable, not using them is leaving money on the table.
Autoscaling is not a substitute for capacity planning. It reacts to load; it does not predict it.
Reserved instances and savings plans reward commitment. If your workload is stable, not using them is leaving money on the table.
Observability is the difference between knowing something is broken and knowing why. Invest in it before you need it.
Spot instances are free money if your workload can tolerate interruption. Most batch and async workloads can.
The shared responsibility model means the cloud secures the infrastructure and you secure everything else. Most breaches are in the everything else.
Backups are not disaster recovery. Test your restores or you do not have backups.
Cloud cost is the new technical debt. Teams that do not track spend by service and team end up with bills that nobody can explain and nobody can reduce.
Backups are not disaster recovery. Test your restores or you do not have backups.
Practical takeaways
- The cloud is just someone else's computer, and that someone else is very good at charging you for convenience.
- Spot instances are free money if your workload can tolerate interruption.
- Reserved instances and savings plans reward commitment.
- IAM is the foundation of cloud security.
- Reserved instances and savings plans reward commitment.
- Reserved instances and savings plans reward commitment.
- Infrastructure as code is not optional at scale.
- The shared responsibility model means the cloud secures the infrastructure and you secure everything else.
- Managed databases trade operational burden for cost and flexibility.
- Multi-region is not multi-cloud.
Infrastructure as code is not optional at scale. Clicking in the console is fine for learning and unacceptable for production.
Spot instances are free money if your workload can tolerate interruption. Most batch and async workloads can.
The shared responsibility model means the cloud secures the infrastructure and you secure everything else. Most breaches are in the everything else.
IAM is the foundation of cloud security. Get it wrong and nothing else matters. Get it right and most attacks become much harder.
Backups are not disaster recovery. Test your restores or you do not have backups.
The cloud is just someone else's computer, and that someone else is very good at charging you for convenience. The question is always whether the convenience is worth the markup.
IAM is the foundation of cloud security. Get it wrong and nothing else matters. Get it right and most attacks become much harder.
Multi-region is not multi-cloud. Running in two AWS regions is redundancy. Running on AWS and GCP is strategy, and it is expensive.
Networking is where cloud bills hide. Data transfer charges between regions and out to the internet add up fast.
Spot instances are free money if your workload can tolerate interruption. Most batch and async workloads can.
IAM is the foundation of cloud security. Get it wrong and nothing else matters. Get it right and most attacks become much harder.
Reserved instances and savings plans reward commitment. If your workload is stable, not using them is leaving money on the table.
Networking is where cloud bills hide. Data transfer charges between regions and out to the internet add up fast.
IAM is the foundation of cloud security. Get it wrong and nothing else matters. Get it right and most attacks become much harder.
Tagging is boring and essential. Without consistent tags, cost attribution is guesswork.
Spot instances are free money if your workload can tolerate interruption. Most batch and async workloads can.
The cloud is just someone else's computer, and that someone else is very good at charging you for convenience. The question is always whether the convenience is worth the markup.
Spot instances are free money if your workload can tolerate interruption. Most batch and async workloads can.
Networking is where cloud bills hide. Data transfer charges between regions and out to the internet add up fast.
Reserved instances and savings plans reward commitment. If your workload is stable, not using them is leaving money on the table.
Sources
This post is informed by discussion on aws.amazon.com and related threads on Hacker News and Reddit. Primary source: https://aws.amazon.com/cloudfront/.
Infrastructure as code is not optional at scale. Clicking in the console is fine for learning and unacceptable for production.
Spot instances are free money if your workload can tolerate interruption. Most batch and async workloads can.
Multi-region is not multi-cloud. Running in two AWS regions is redundancy. Running on AWS and GCP is strategy, and it is expensive.
Spot instances are free money if your workload can tolerate interruption. Most batch and async workloads can.
Autoscaling is not a substitute for capacity planning. It reacts to load; it does not predict it.
IAM is the foundation of cloud security. Get it wrong and nothing else matters. Get it right and most attacks become much harder.
Reserved instances and savings plans reward commitment. If your workload is stable, not using them is leaving money on the table.
Backups are not disaster recovery. Test your restores or you do not have backups.
Serverless is great until it isn't. Cold starts, concurrency limits, and vendor lock-in are real costs that show up at scale.
Cloud cost is the new technical debt. Teams that do not track spend by service and team end up with bills that nobody can explain and nobody can reduce.
Managed databases trade operational burden for cost and flexibility. Know what you are giving up.
Autoscaling is not a substitute for capacity planning. It reacts to load; it does not predict it.
Observability is the difference between knowing something is broken and knowing why. Invest in it before you need it.
Multi-region is not multi-cloud. Running in two AWS regions is redundancy. Running on AWS and GCP is strategy, and it is expensive.
Autoscaling is not a substitute for capacity planning. It reacts to load; it does not predict it.
Multi-region is not multi-cloud. Running in two AWS regions is redundancy. Running on AWS and GCP is strategy, and it is expensive.
The cloud is just someone else's computer, and that someone else is very good at charging you for convenience. The question is always whether the convenience is worth the markup.
Serverless is great until it isn't. Cold starts, concurrency limits, and vendor lock-in are real costs that show up at scale.
Reserved instances and savings plans reward commitment. If your workload is stable, not using them is leaving money on the table.
IAM is the foundation of cloud security. Get it wrong and nothing else matters. Get it right and most attacks become much harder.
IAM is the foundation of cloud security. Get it wrong and nothing else matters. Get it right and most attacks become much harder.
Networking is where cloud bills hide. Data transfer charges between regions and out to the internet add up fast.
Infrastructure as code is not optional at scale. Clicking in the console is fine for learning and unacceptable for production.
The cloud is just someone else's computer, and that someone else is very good at charging you for convenience. The question is always whether the convenience is worth the markup.
Spot instances are free money if your workload can tolerate interruption. Most batch and async workloads can.
The shared responsibility model means the cloud secures the infrastructure and you secure everything else. Most breaches are in the everything else.
Multi-region is not multi-cloud. Running in two AWS regions is redundancy. Running on AWS and GCP is strategy, and it is expensive.
Autoscaling is not a substitute for capacity planning. It reacts to load; it does not predict it.
Spot instances are free money if your workload can tolerate interruption. Most batch and async workloads can.
Reserved instances and savings plans reward commitment. If your workload is stable, not using them is leaving money on the table.
Managed databases trade operational burden for cost and flexibility. Know what you are giving up.
Infrastructure as code is not optional at scale. Clicking in the console is fine for learning and unacceptable for production.
Networking is where cloud bills hide. Data transfer charges between regions and out to the internet add up fast.
Backups are not disaster recovery. Test your restores or you do not have backups.
Spot instances are free money if your workload can tolerate interruption. Most batch and async workloads can.
The cloud is just someone else's computer, and that someone else is very good at charging you for convenience. The question is always whether the convenience is worth the markup.
Managed databases trade operational burden for cost and flexibility. Know what you are giving up.
Networking is where cloud bills hide. Data transfer charges between regions and out to the internet add up fast.
The cloud is just someone else's computer, and that someone else is very good at charging you for convenience. The question is always whether the convenience is worth the markup.
Managed databases trade operational burden for cost and flexibility. Know what you are giving up.
Reserved instances and savings plans reward commitment. If your workload is stable, not using them is leaving money on the table.
Backups are not disaster recovery. Test your restores or you do not have backups.
Reserved instances and savings plans reward commitment. If your workload is stable, not using them is leaving money on the table.
Reserved instances and savings plans reward commitment. If your workload is stable, not using them is leaving money on the table.
Spot instances are free money if your workload can tolerate interruption. Most batch and async workloads can.
The cloud is just someone else's computer, and that someone else is very good at charging you for convenience. The question is always whether the convenience is worth the markup.
Cloud cost is the new technical debt. Teams that do not track spend by service and team end up with bills that nobody can explain and nobody can reduce.
Autoscaling is not a substitute for capacity planning. It reacts to load; it does not predict it.
Spot instances are free money if your workload can tolerate interruption. Most batch and async workloads can.
Networking is where cloud bills hide. Data transfer charges between regions and out to the internet add up fast.
Observability is the difference between knowing something is broken and knowing why. Invest in it before you need it.
Reserved instances and savings plans reward commitment. If your workload is stable, not using them is leaving money on the table.
Infrastructure as code is not optional at scale. Clicking in the console is fine for learning and unacceptable for production.
Backups are not disaster recovery. Test your restores or you do not have backups.
Autoscaling is not a substitute for capacity planning. It reacts to load; it does not predict it.
Managed databases trade operational burden for cost and flexibility. Know what you are giving up.
Multi-region is not multi-cloud. Running in two AWS regions is redundancy. Running on AWS and GCP is strategy, and it is expensive.
Tagging is boring and essential. Without consistent tags, cost attribution is guesswork.
Tagging is boring and essential. Without consistent tags, cost attribution is guesswork.
Autoscaling is not a substitute for capacity planning. It reacts to load; it does not predict it.