Changelog #0162

Welcome to the cloud! We got fun and games. We got nearly everything you want. Buddy, we know the names!

The word “nearly” is doing a lot of work here — because we do have some explaining to do.

Railway experienced a major outage today (Friday) prompting the whole team to give extra special attention to our customers.

In addition to publishing a detailed incident report and a deeper editorial blogpost on the incident, we also alerted affected customers across the Community Forum, Email, Twitter, and Discord.

We’ll return to the incident in a sec, but let’s first jump into the rest of this week’s changelog.

Last week we announced that our company was pivoting into becoming a social deployments company.

audience booing

Just kidding. But we did launch usernames. Now we’re talking.

Whether in the Community Forum or in the Template Library, usernames give you another dimension to express yourself. Claim yours now!

Just visit https://railway.app/account and set one up for yourself.

We’ll also be requiring usernames for the forum moving forward, or we’ll be happy to provide one to you. (This means you, user-9u297402.)

While we are talking about the forum, Engineer Ray folded in ALL of the Canny upvotes and feature requests into the community.

We invite you to play Lawyer for your favorite feature(s) … your client just might win!

We look forward to seeing your suggestions get promoted to our product roadmap.

First off, I don’t know what other companies say, but we have the best moderators and maintainers in the world and we owe an extra special thanks to the Conductors for bringing this up…

We now support variable chain references!

This means that if you assign the following variables:

# Service A
VAR_A=${{service-b.VAR_B}}

# Service B
VAR_B=${{service-c.VAR_C}}

# Service C
VAR_C=hello

The environment variables will now resolve to:

VAR_A=hello
VAR_B=hello
VAR_C=hello

This is useful for times that you only want to source variables from a service and not from the global scope. This is also a pre-requisite for something really cool that the writer of this week’s changelog isn’t allowed to leak…

Stay tuned!

Now on to the serious stuff.

It’s our policy to share the public details of every production outage and today we experienced a thorny one that impacted 30% of the workloads on US-West-1.

As mentioned above, we wrote an extremely detailed incident report here.

And Railway CEO Jake wrote an editorial blogpost on the outage here, so check that out if you’d like some more context on what happened and why.

Members of the Customer Success and Support teams also communicated the impact to affected customers across the Community Forum, Email, Twitter, and Discord.

For the future, we highly recommend following @railway_status on Twitter for the latest in all service disruptions.

Thanks for your support.

That’s all for this week, enjoy your weekend!