Ctrip's Confluence data center operation practice with more than 30,000 users

This article comes from Atlassian customer stories. It tells about my experience as a system administrator in operating a Confluence data center for more than 30,000 employees. I hope it will be helpful to the community.

Original: https://www.atlassian.com/zh/customers/ctrip

QQ截图20210616154150.png

Ctrip started using Confluence in 2013. At that time, it mainly served the R&D team for demand management, meeting minutes, and collaborative editing. Between 2013 and 2019, Ctrip's business and functional departments have their own tools similar to Confluence, and some teams also use Confluence Cloud, but the data of each department is completely separated from each other, and everyone does not have a unified platform for knowledge Management and information islands are more serious.

The technical team tried to push Confluence to the entire company in 2017, but the server version at that time only supported the technical team, the performance was already tight, especially Ctrip made some customizations on the Confluence homepage. Many pictures and menu displays need to determine user permissions, but Ctrip has a large number of user groups, tens of thousands, and the user just logs in to the home page for this operation will have a waiting time of 5-20 seconds, and performance is not satisfactory. Conditions for company-wide promotion.

In order to improve the usability of the system, in 2019, we upgraded the original Server version to Confluence Data Center and started to promote it to the whole group. We used Confluence as a company-wide knowledge base platform. We mobilized the human resources department of the whole group. More than 30 HR students from various business units were invited as space managers to join Confluence's promotion work. Now the whole group has used Confluence as a unified knowledge base.

After the Confluence Data Center version was officially launched at the end of 2019, the number of users in the month increased from the original more than 6,000 users to more than 15,000 users. We conducted four consecutive six-month promotions within the group to promote all departments of the group Start using Confluence in full. The result of the third NPS survey was a net recommendation value of 55%. Among the thousands of internal users interviewed, Confluence's satisfaction score was 9 out of 10!

Here are some real feedback from front-line employees on Confluence:

Ctrip's administrative department: "Confluence's entire company's high-quality documents are centralized, updated in a timely manner, and quickly synchronized, so that we can understand the company's latest developments."

Ctrip's operating department: "Compared with the network disk, Confluence can build a page framework, personalized editing, and easier to read."

Ctrip HR department: "Confluence multiple people jointly maintain knowledge points, help the team complete knowledge sharing, and unearth in-depth knowledge contributors."

Ctrip customer service department: "Confluence is organized and structured to provide a window for new team members to quickly understand internal communication and collaboration mechanisms."

 

How can tens of thousands of employees quickly use Confluence Data Center?

In the process of guiding internal personnel to use Confluence, in addition to the official operation guide, we provided users with common problems and solutions in both Chinese and English, and recorded operation videos (20 videos for ordinary users, space administrators, and system administrators, A video of about 6 minutes, the shortest possible way allows users to quickly locate the function and operation method they want to know);

We only have 2 part-time system administrators, and we cannot support tens of thousands of users. The administrator must do a good job in the training and authorization of the space administrator, and the space administrator should maintain his own site.

The administrator needs to draw up Confluence usage rules, such as attachment size limit of 50MB, and the user explains that Confluence is not a storage tool, and large files are stored on the network disk according to the company's requirements.

In addition, the administrator must remember to archive the discarded space regularly.

As a system administrator, what should I do if I encounter a problem?

First of all, for system-level issues (such as errors, performance issues, etc.), it is recommended to consult Atlassian certified partners in China, or log in to the original factory community or service desk for consultation;

Second, you can join the WeChat community, and more than 2,400 Chinese regional partners will discuss best practices and solutions with you;

Finally, if you are a Data Center user, Atlassian's official account manager currently forms a WeChat enterprise user group in mainland China, and you will connect regularly every month.

 

How to better upgrade Confluence Data Center?

The Confluence Data Center version adopts an annual subscription model. Users can choose to upgrade to the latest version at any time. The administrator must consider when to upgrade, how to choose the frequency and version of the upgrade, and other issues.

In Ctrip Group, we choose to perform a version upgrade in about 1.5 to 2 years. The latest LTS version is preferred for upgrade. The version upgrade plan is planned in advance, and partners are consulted. In addition, we can pay attention to the official key functions in each version of Release Note. And the bugs that have been publicized in the current version.

In addition to the stability of the version, we will try our best to choose a version that contains more new features when upgrading. For example, the current 7.11 version integrates the original 2 plug-ins: website branch report, team calendar, etc.

From Server to Data Center version, as a Confluence administrator, the most satisfying feature is a multi-node cluster. Even if a certain node is down, the system can still operate normally, and you can even restart the nodes in turn during working hours. At the same time, ensure that the system is continuously available.

 

Tips:

With the Data Center version, except for the shutdown during the upgrade, the system is always available for the rest of the time;

Using the Data Center version, occasional network fluctuations or communication interruptions can be solved by restarting the single nodes one by one;

With the Data Center version, every 1-2 months, a single node restarts in turn, and the page opening speed will be improved.

How to better improve the system security of Confluence Data Center?

As a system administrator, there is a problem that must be faced: how to restrict users from downloading attachments and how to ensure that information is not leaked?

Regarding this problem, whether it is the use of plug-ins, third-party system monitoring or other measures, we must first make it clear that everything we do must ensure the company's information security without affecting the normal use of users.

Whether it is watermarking or other means, the purpose is not to pursue accountability after an information leak occurs. Of course, this is just a bottom line measure.

Through the propaganda of the internal information security rules of the enterprise, we let users know what can be done and what cannot be done, which ones are the red lines of safety, and what are the consequences after crossing the line. These need to be clearly publicized and regularly promoted within the organization. . At the same time, some internal technical control measures, such as network monitoring, prohibition of external USB drives, whitelisting of outgoing emails, and sensitive information scanning, are used to protect the security of corporate data.

As of the end of 2020, Confluence has been fully used in all functional departments of the Ctrip Group. From the initial R&D team to the current group-wide business team, Confluence provides us with a stable and high-performance cross-departmental collaboration platform:

More than 30,000 users visit Confluence every month, and the Data Center version has reached 25 million page views in one year;

Add about 100,000 pages and attachments every month, generating 300,000 page updates;

The space classification includes company strategy, rules and regulations, knowledge inheritance, technology sector, knowledge map, etc., with a total of about 140+ independent spaces.

 

2 comments

Alex Medved _ConfiForms_
Rising Star
Rising Star
Rising Stars are recognized for providing high-quality answers to other users. Rising Stars receive a certificate of achievement and are on the path to becoming Community Leaders.
June 16, 2021

Hi

Great article, thanks for sharing!

How do you manage user permissions (user groups) and clean ups (personal spaces of removed/disabled users, for example)?

Also, do you track how much space each user uses (personal spaces, attachments) - do you have some monitoring tools that help you with that (any quota enforcement maybe)? I understand you have limited the size of the attachment and try to educate your users, but still, no one prevents them from uploading multiple files of 50MB each...

Alex

Ollie Guan
Community Leader
Community Leader
Community Leaders are connectors, ambassadors, and mentors. On the online community, they serve as thought leaders, product experts, and moderators.
June 16, 2021

Hi @Alex Medved _ConfiForms_ ,

We use LDAP to manage users and user groups, sort out and archive the personal space of departed users once a quarter.

We will not monitor how many attachments users upload. At present, the total storage of attachments is still under control, and most of them are pictures, Word, Excel, PPT, PDF, etc.

Regarding attachments, I think the best way to guide users is not to set restrictions, but to clearly inform users that attachments of different specifications and formats should be stored in appropriate tools. Of course, we also need to provide various storage tools for our classification, such as network Hard disk, version control tools, etc.

Comment

Log in or Sign up to comment
TAGS
AUG Leaders

Atlassian Community Events