Skip to content

Commit

Permalink
Vitess user survey blog (#1832)
Browse files Browse the repository at this point in the history
* user survey blog init

Signed-off-by: kirtanchandak <themarketingkirtan@gmail.com>

* complete blog added

Signed-off-by: kirtanchandak <themarketingkirtan@gmail.com>

* formatting fix

Signed-off-by: kirtanchandak <themarketingkirtan@gmail.com>

---------

Signed-off-by: kirtanchandak <themarketingkirtan@gmail.com>
  • Loading branch information
kirtanchandak authored Aug 28, 2024
1 parent 4455ef1 commit 90544a4
Show file tree
Hide file tree
Showing 16 changed files with 140 additions and 0 deletions.
140 changes: 140 additions & 0 deletions content/en/blog/2024-08-27-vitess-user-survey.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,140 @@
---
author: 'Kirtan Chandak'
date: 2024-08-27
slug: '2024-08-27-vitess-user-survey'
tags: ['Vitess', 'Sharding', 'MySQL']
title: 'Vitess User Survey'
description: "Summary responses and insights from Vitess user survey."
---

We recently conducted a survey of how Vitess is being used by the community. This blog post summarizes what we learned.

> **"Vitess solves an existential threat for services which outgrow a single MySQL database."**
> **"Horizontal sharding helps us scale quickly, and the new generation execution plan, Gen4, enables us to support more SQL queries."**
> **"Vitess has allowed us to scale and step away from our dev-ops role, allowing us to focus on higher level and higher impact tooling and automation."**
# Background

The story begins with the Vitess maintainer team wanting to conduct a user survey with community members and engineers using Vitess in production to gain a better understanding of how to improve the project. As an LFX Mentee'24, I worked with the Vitess maintainers to form the questions and create a format for conducting the user survey. We drafted the questions, reviewed them, and developed a final version of the survey. I'm happy to report that we received a number of responses.

# Summary Responses

### 1. How long have you been using Vitess?
First of all, we decided to ask how long our community members have been using Vitess. Interestingly, quite a few people have been using Vitess for more than 3 years.
<div style="width:500px">
<img src="/files/2024-08-27-vitess-user-survey/1-vitess-long.png" alt="Distribution: 45% More than 3 years, 36% - 1 to 3 years, 13% - 6 months to 1 year, and 6% - Less than 6 months." style="margin-top:6px"/>
</div>

### 2. Are you running Vitess in production?
More than half of Vitess users have been in production for over 3 years, and an additional 25% have been in production for over 1 year.
<div style="width:500px">
<img src="/files/2024-08-27-vitess-user-survey/2-production.png" alt="Distribution: 1 to 3 years at 50%, 6 months to 1 year at 27%, Yes for less than 1 year at 14%, and More than 3 years at 9%" style="margin-top:6px"/>
</div>

### 3. How often do you upgrade your Vitess version?
The most popular frequency for Vitess version upgrades is once a year. New versions of Vitess are released 3 times a year, and many users follow the same schedule.
<div style="width:500px">
<img src="/files/2024-08-27-vitess-user-survey/3-vitess-version-often.png" alt="Distribution: More than 2 years (27%), Once a year (27%), Every 1-2 years (23%), No fixed cadence (13%), and Every 6 months (9%)" style="margin-top:6px"/>
</div>

### 4. How often do you upgrade your MySQL version, including 8.0 minor version upgrades?
The most popular frequency for upgrading MySQL version is once a year. Most people upgrade their MySQL version at least once every two years.
<div style="width:500px">
<img src="/files/2024-08-27-vitess-user-survey/4-minor.png" alt="Distribution: More than 2 years between upgrades and Once a year have 27%, Every 1-2 years has 23%, No fixed cadence has 13%, and Every 6 months has 9%" style="margin-top:6px"/>
</div>

### 5. Which version of MySQL are you running?
The most popular MySQL version among our respondents is 8.0, with a slightly higher proportion of users on Percona vs Oracle MySQL.
<div style="width:500px">
<img src="/files/2024-08-27-vitess-user-survey/5-current-version.png" alt="Distribution: Percona 8.0 leads with 45%, Oracle 8.0 has 36%, Percona 5.7 shows 14%, and Mariadb 10.11 is the least used with 4%" style="margin-top:6px"/>
</div>

### 6. Which version of MySQL do you plan to upgrade to next?
More than half of users have responded with NA and TBD for their next MySQL version upgrade. Others are planning to upgrade to Percona 8.x and Oracle 8.x in the future.
<div style="width:500px">
<img src="/files/2024-08-27-vitess-user-survey/6-next.png" alt="Distribution: NA (32%), TBD (27%), Percona 8.x (22%), Oracle 8.x (19%)" style="margin-top:6px"/>
</div>

### 7. How soon do you plan to upgrade to the next MySQL version?
A majority of our users plan to perform a MySQL version upgrade sometime in the next two years.
<div style="width:500px;">
<img src="/files/2024-08-27-vitess-user-survey/7-soon.png" alt="Distribution: 6 months to 1 year (45%), TBD (36%), More than 2 years (14%), and 1 to 2 years (4%)" style="margin-top:6px"/>
</div>

### 8. How do you deploy Vitess?
Users deploying Vitess on Kubernetes are in the majority, followed by bare metal, VMs, etc. Interestingly, some deployments use a mixed architecture.
<div style="width:500px;">
<img src="/files/2024-08-27-vitess-user-survey/8-deploy.png" alt="Distribution: Kubernetes leads with 54%, followed by Mix of Kubernetes, VMs, bare metal, VMs, and Bare Metal at 14% each, others account for 4%." style="margin-top:6px"/>
</div>

### 9. How big is your data size?
A significant majority of Vitess deployments have a data size greater than 10 TB.
<div style="width:500px;">
<img src="/files/2024-08-27-vitess-user-survey/9-data.png" alt="Distribution: More than 10 TB: 73%, 1 TB to 10 TB: 9%, 250 GB to 1 TB: 9%, less than 250 GB: 9%" style="margin-top:6px"/>
</div>

### 10. How big is your Vitess deployment in terms of keyspaces?
Most Vitess users have fewer than 100 keyspaces, but some have many more.
<div style="width:500px;">
<img src="/files/2024-08-27-vitess-user-survey/10-keyspaces.png" alt="Distribution: 1-100' with 77%, '100-1000' with 9%, '1' with 9%, and 'Greater than 1000' with 4%" style="margin-top:6px"/>
</div>

### 11. How big is your Vitess deployment in terms of shards?
Most Vitess users have fewer than 100 shards, but a significant number of them have a high number of shards.
<div style="width:500px;">
<img src="/files/2024-08-27-vitess-user-survey/11-shards.png" alt="Distribution: 1-100 is 59%, 100-1000 is 18%, 1 is 18%, and Greater than 1000 is 5%" style="margin-top:6px"/>
</div>

### 12. Have you run a resharding process?
A majority of the Vitess users have gone through a resharding process.
<div style="width:500px;">
<img src="/files/2024-08-27-vitess-user-survey/12-resharding.png" alt="Distribution: Yes is 77% and is longer than the bottom bar labeled No which is 23%" style="margin-top:6px"/>
</div>

### 13. Do you contribute code to Vitess?
50% of our respondents contribute code to Vitess.
<div style="width:500px;">
<img src="/files/2024-08-27-vitess-user-survey/13-contribute.png" alt="Distribution: The top bar labeled Yes shows 54%, and the bottom bar labeled No shows 46%" style="margin-top:6px"/>
</div>

### 14. Cloud
50% of our users use Public cloud and most of them use GCP and Azure.
<div style="width:500px;">
<img src="/files/2024-08-27-vitess-user-survey/14-cloud.png" alt="Distribution: Public: 50%, On Prem: 33%, Private: 17%, GCP: 42%, AWS: 33%, and Azure: 25%" style="margin-top:6px"/>
</div>

### 15. Which major version of Vitess are you running?
23% of the users are running version 19.0 and other versions has a distributed split as seen in the graph below.
<div style="width:500px;">
<img src="/files/2024-08-27-vitess-user-survey/15-version.png" alt="Distribution: 19.0: 23%, 18.0: 18%, 17.0: 14%, 16.0: 14%, 15.0: 9%, 11.0: 9%, 20.0: 9%, 14.0: 4%" style="margin-top:6px"/>
</div>

# Insights

### We asked our users how Vitess has helped their business. We got some interesting answers from different companies.

1. Scalability: Many Vitess users have scaled the performance of their databases using horizontal sharding.
For example:
[How Shopify horizontally scaled the backend of an important Rails application.]("https://shopify.engineering/horizontally-scaling-the-rails-backend-of-shop-app-with-vitess")

2. Operational Efficiency: Vitess reduced the burden of managing MySQL clusters at scale, especially in distributed environments like Kubernetes.
3. High Availability: Hardware issues do not cause downtime while they are being repaired.
4. High Connections: Vitess is able to manage high connection counts and rates of churn, essential for highly trafficked and concurrent applications.
5. Efficient Online Schema Changes: Online DDL support in Vitess, combined with horizontal sharding, has let users efficiently change their schema, reducing operational workload for faster product iteration.
6. Lowered DevOps Burden: By abstracting away the complexities of sharding and scaling, Vitess empowers teams to concentrate on higher-level development rather than getting bogged down in database management.
7. Flexibility and stability: Vitess has brought stability even to unsharded setups by managing the connection limits and setting the number of concurrent threads so that databases run as smoothly as possible without any performance issues.

### We also asked them what the biggest missing feature in Vitess is. Here is a curated sample of the responses.

1. Improved Observability: A consistent dashboard for observing the vreplication streams, with controls over schema changes, can give a better overview of operations.
2. Unified Web UI: A single, easy-to-use web interface for all administrative tasks would ease management and enhance usability.
3. Circuit Breaking and Adaptive Concurrency Limits: We need features to prevent clients from being blocked by slow shards, and manage concurrency.
4. More Clear Upgrade Path with LTS Versions: a well-defined upgrade path with LTS versions is important for stability and fewer bugs.
5. Auto Sharding Improvements: More advanced auto-sharding capabilities would be nice, such as in TiDB.
6. Vitess Keyspace History: Add a feature to track and review changes over time to a keyspace -- such as reparents, backups, or scaling events for historical analysis and debugging.

# Conclusion

The maintainer team appreciates your responses to the survey. Your feedback will be incorporated into the project roadmap.
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.

0 comments on commit 90544a4

Please sign in to comment.