Agreed! We have another episode coming up with Kate + another AWS databases SME coming up where we cover RDS optimization. Stay tuned :)

Appreciate you for listening! If there are other topics you'd like to see us cover in a future episode, let me know here :)

Exactly - the simplification mindset totally applies both ways! As Kate mentioned, AWS deliberately (we think, at least) stripped down DSQL to push developers toward simpler, cleaner database patterns. For apps that need massive scale but can live without fancy DB features, going minimal makes a lot of sense.

Hey there! This is a very common scenario that we see across our customer base. Here are some things to look for in your database and application: > Review performance insights for your top queries, perform `EXPLAIN ANALYZE ` on the queries and compare the query plans between your test environment and production >> Be sure to review your database code for any index hints, query plans should be reviewed with and without index hints. >> Review your sql statements for large `IN (....) ` predicates, refactor these queries to use a different approach (eg: JOIN, temp table, etc) > Make sure transactions are as short as possible, and that your application is not holding them open for longer than necessary > Batch transactions together where possible > Review/implement the Amazon Aurora enhanced binlog feature, this can significantly reduce CPU but be aware of the limitations of this feature >>aws.amazon.com/blogs/database/introducing-amazon-aurora-mysql-enhanced-binary-log-binlog/ >> docs.aws.amazon.com/AmazonRDS/latest/AuroraUserGuide/AuroraMySQL.Enhanced.binlog.html#AuroraMySQL.Enhanced.binlog.limitations >> AWS DMS is not compatible with enhanced binlog > Review custom parameters and make sure they are still relevant > Review your application's connection methodology. >> Is connection pooling in use? >> Often, a spike in connections is due to queries running for longer, creating a backlog of tasks and an increase in connections >> is the database driver up to date? > Review open bugs for your MySQL version and see if any apply to you. >> bugs.mysql.com/search.php?search_for=&status=Active&severity=&limit=10&order_by=&cmd=display&direction=ASC&bug_type%5B%5D=Server%3A+Optimizer&os=0&phpver=8.0.32&bug_age=0 If you'd like to book a call with our team to dive deeper into this use case, please see www.doit.com/ and "Book a call"

When you compare apples to apples, Fastly and CloudFront have very similar characteristics. Additionally, when the origin of the files is S3, you need to account for the Data Transfer Out cost from AWS to Fastly. When using CloudFront, this cost is $0.

Piggy-backing off of @avikeinan8616 - Here's a clip from the episode of him describing the savings in data transfer costs when serving data from S3 through CloudFront vs. a 3rd-party CDN: kzbin.infomBF71w00IO0?feature=share

@@avikeinan8616 Understood that of course you pay twice with a non-AWS CDN. For well cachable objects the dominant cost will be the CDN bandwidth, not the origin one. Plus Fastly has a much better feature set.

Hi @JoshRenton91, glad you're finding the videos helpful :) For an Amazon EBS volume, the determination of whether I/O operations are sequential or random is not about the timing of the operations, but rather about the pattern in which data is accessed. Sequential I/O operations typically involve accessing contiguous blocks of data on the storage device.This means that the data blocks are located next to each other on the storage device. On the other hand, random I/O operations occur when data is read or written in a non-continuous, scattered manner. This means that the data blocks are not located next to each other on the storage device.

@@CiaraMcMahon-p8f Thank you so Ciara, forgive my ignorance, but how would one determine whether one is writing to contiguous blocks or not? If I instruct my application to write a file to disk, how will I know if it has written contiguously or not? And then in the context of EBS, how would I know if EBS as stored the blocks contiguously?

Yep! Here's the link to the full presentation: kzbin.info/www/bejne/hILHdaZ-r66ZmtE Sorry it took so long - had to dig it up as we recorded this back in late-2023.

Thanks for your patience! Here is the repo with the sample code: github.com/doitintl/genai-kyc

Yep: kzbin.info/www/bejne/hajQqWdmpJh6odksi=QV5V2pHfoI-W-kqQ Also check out the episode description for some additional resources mentioned during the full episode :)

Appreciate your kind comment! Enjoy the rest of the series :)

Could be a good idea! Can you elaborate a bit more? And which terms that we used this video did you have in mind?

😂

Thanks for watching/listening|! Anything specific you'd like us to dive deeper into?

As of July 5th 2023, you can't switch a project to be on flat-rate pricing, so your options for BigQuery are either on-demand or Editions. Some follow-up questions to help us make a recommendation: Are you consistently using 2000 slots during 4am - 12 noon, or is it very spiky? If spiky, are the spikes at the same times during that timeframe? Additionally, how many TiB of data is being processed per day during that timeframe? Last question, are you using anything like Looker or dbt? If so, there could be some tricks you could do to bring your slot consumption down from 2000.

Thanks for watching the episode + glad to hear you liked the content :). Now to answer your question (via Kate from the episode): "MySQL is in general a single-threaded database designed for high volumes of queries as opposed to large analytic or data warehouse type queries. MariaDB and MySQL are optimized for different types of workloads. Something that works well in one may not work well in the other. MySQL tends to excel at complex joins and sub-queries whereas MariaDB is often better at recursive CTEs and analytics. If you wished to stay using MySQL rather than MariaDB, then you can look at parallel query on RDS Aurora MySQL. This allows you to run complex queries directly against the storage layer in a parallel fashion. docs.aws.amazon.com/AmazonRDS/latest/AuroraUserGuide/aurora-mysql-parallel-query.html"

@@doitintl Thank you, perhaps we will give it a try

Here you go: engineering.doit.com/aws-public-ipv4-price-increase-the-complete-guide-73c4c62ae18e I've also just updated the episode description with it, so thanks for asking!

AlterNAT - github.com/1debit/alternat

Glad you enjoyed