Skip to main content

Hartwork Blog

Free Software, Music, Chinese Chess

Uptime monitoring with Healthchecks.io (is possible!)

Besides this blog, I operate a small set of web services:

Until recently, I have been monitoring their availability using Cabot. Cabot is Open Source, self-hostable, written in Python, based on Django, sends e-mail when things start to appear offline — not too bad. But its maker stopped maintaining it over four years ago and it is still using Python 2.7; forking the project was not an option to me, so it was time to look for alternatives to Cabot.

I remembered Healthchecks(.io) from briefly seeing it in the past but it never came up in the results when googling for "uptime monitoring" and page Healthchecks.io Compared to Cronitor eventually made it clear — website monitoring is not supported.

Healthchecks.io does not support uptime monitoring of websites

But if that's the answer — am I asking the right question?

What if the task of uptime monitoring — sending e-mail when things go offline — could be solved with Healthchecks but not inside of it?

It turned out that all it takes is a small external ingerient: a small tool checking the websites frequently and reporting to Healthchecks about whether that worked: Healthchecks would treat it like any other cron job and alert me whenever pings would stop coming in — cool!

Long story short: For that added piece I ended up writing a simple container cron — pytocron — with native Healthchecks integration. Given a crontab file like this…

# hc-ping: https://hc-ping.com/xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx
1/30 * * * * * * online https://blog.hartwork.org/ 'Hartwork Blog'

# hc-ping: https://hc-ping.com/yyyyyyyy-yyyy-yyyy-yyyy-yyyyyyyyyyyy
2/30 * * * * * * online https://jawanndenn.de/ 'Create a new poll'

# hc-ping: https://hc-ping.com/zzzzzzzz-zzzz-zzzz-zzzz-zzzzzzzzzzzz
3/30 * * * * * * online https://linuximages.de/ 'openstack'

…pytocron will now check these three websites every 30 seconds and use the URLs from the # hc-ping: <URL> comment lines to report to Healthchecks whether these checks succeeded. Just one small container more in my herd.

Command online in this is /usr/local/bin/online and this simple custom-made script:

#! /usr/bin/env bash
# Copyright (c) 2025 Sebastian Pipping <sebastian@pipping.org>
# SPDX-License-Identifier: 0BSD

set -e -u

if [[ "${#}" -ne 2 ]]; then
    echo "usage: online URL NEEDLE" >&2
    exit 1
fi

url="${1}"
needle="${2}"

wget -qO- -T3 "${url}" | grep -qF "${needle}"

Combined with the Healthchecks dashboard, I even have a simple status page now at status.pipping.org.

That's all it takes to do uptime monitoring with Healthchecks.

Was that interesting and/or helpful to you? Please let me know.

Best, Sebastian

DomainFactory WTF 2025

Or how DomainFactory — part of Host Europe and a GoDaddy company — takes turning against their customers to a whole new level with their 2025 forced e-mail migration to Microsoft Office 365, and breaks things left and right in the process…

Some background before we start

DomainFactory had an excellent reputation 20 years ago. I was one of those new customers that they won through an exciting offer for their 5th anniversary in 2005. Everything times five: 5 domains, 5 GB webspace, 5 databases and so on — good times, and good usability. I was a happy customer.

Then in 2015(?) the company announced to move their data centers out of Germany to Strasbourg in France. I can only guess that the amount of protest about that out-of-Germany move got DomainFactory to eventually only move some of their servers to Strasbourg and to offer hosting in Germany as an alternative, for a premium. So customers had been hosted in Germany for years and suddenly were charged a premium for what they already had for free before — cool! I'm pretty sure that's illegal but you tell me. I have not forgotten or forgiven that move and it was the point I stopped recommending DomainFactory to others. A similar thing happened again later when they stopped offering that 5x-everything deal to existing customers and force-migrated me to a different deal that was more expensive; nothing in that change was of my interest.

Now in 2025 things changed to a whole new level of WTF.

The forced migration of e-mail to Office 365 in March/April 2025 — what a disaster!

I will save you and me a timeline-like view of the story and instead just focus on the damage that DomainFactory has done:

  • Three days before the start of the migration, access to the mailbox section of the admin panel started to be denied altogether. So I had no way of seeing the list of mailboxes or deleting mailboxes so that their contained e-mails would not be copied to Microsoft. DomainFactory's note "[W]e will freeze your mailbox settings three days before the scheduled migration." (translated from German) in their announcement e-mail did not stand out, and was easy to miss unless you read every single word of that e-mail.

  • Two of three mailboxes affected by migration had >100 items skipped, and migration was stopped and not finished. DomainFactory did not inform or apologize or communicate plans about how to continue. How do I know? From my own search in the Exchange admin center they expose labeled "Advanced". That's the same mailboxes will start to be charged 0.99 EUR per month per mailbox (from April 18th for me) when they have been included for free for years. With corrupted e-mails (and a load of friction in migration) but now paid. And they did say that "No further action is required on your part before the migration." (translated from German) Except for doing backups or jumping ship maybe.

  • My forwarding alias used for my account with PayPal was deleted, putting my use of PayPal at risk; the user interface no longer allows the creation of cross-domain mail aliases which was the case for my forwarder with PayPal.

  • Customer support has closed 3 of 3 support requests created via One.Done! with a template reply and without resolving my problems. They were:

    • The fact that the PayPal forwarder alias was lost in the migration. That issue is unresolved. (DF-748556)
    • The fact that they blocked two domain transfers despite use of the correct Auth-Code. (DF-760730)
    • (The fact that all folders seemed gone after re-login to IMAP. It turned out to need a re-subscription to all folders on my end. Would have been easy to send a helpful response to. (DF-748751))

  • Waiting time for customer support by phone was 40 minutes (on April 3rd, afternoon). My total waiting time to speak with DomainFactory phone support (in the last two weeks) was roughly 40 + 25 + 25 + 7 = 97 minutes.

  • Sending e-mails with Thunderbird on Linux turned out impossible to get to work again after migration. I don't think it's by accident that their official help on Thunderbird says "Mac and Windows" but does not mention Linux. Yes, I did try to enable "app" SMTP auth but without any observable change in log-in failure.

  • Because address sebastian@pipping.org was realized as an alias, the new webmail at outlook.office365.com would not allow me to send e-mail as my primary address sebastian@pipping.org. So sending as sebastian@pipping.org did not work in any way. For each address, the first attempt to log in via webmail would also fail with a weird error that went away after I re-tried with "flagging enabled".

  • At no point did I approve Microsoft's terms of service, yet they now have a copy of all my e-mails. No idea in which jurisdiction or data center. Also, if you had asked me: I wanted zero Microsoft in my e-mail infrastructure — zero. By now access to productivity.secureserver.net is blocked to me unless I accept Microsoft terms: the closing button disappeared for me on April 15th.

  • Deleting of one mailbox ...@hartwork.org (that happened to be the last postbox on that domain) killed all aliases for hartwork.org as a side effect (and e-mail to them was rejected). These aliases magically re-appeared when I dared experiment with the restore deleted users feature and could hardly believe my eyes when that in fact restored aliases unrelated to the mailbox but on the same domain.

  • (One night the whole admin panel was impossible to log into when I was about to adjust settings.)

  • All server-side filters are gone since the migration. Previously, certain mail headers would move e-mails to — say — mailing list bug-make into a dedicated folder, likely based on technology procmail. All gone, everything goes straight to flooding INBOX now.

  • At the time of this writing, I have already lost more than a full week of time and effort in cleaning up this mess. Thanks DomainFactory!

My migration to INWX for domains and DNS and to Manitu Webhosting for e-mail is ongoing still — at 95% maybe —, it takes time. Having recent backups from the use of offlineimap helped a ton. Many thanks to that project!

I did look at providers that target e-mail hosting specifically and in Germany before I understood that I should look for webhosting that includes e-mail instead, both in terms of pricing and in terms of flexibility with creation of new IMAP mailboxes over time. I ruled out…

  • goneo because they insist on also taking over the nameserver for custom domains.
  • mail.de because they do not seem to support serving custom domains as an MX server and need one account per IMAP mailbox.
  • mailbox.org because — longer story (..) — multiple IMAP accounts are not fun to manage and rather expensive with them.
  • Posteo because they decided to not support custom domains (and seem to need one account per IMAP mailbox).
  • Tuta because they deliberately do not support IMAP.

Let me add a few pictures from the crime scene:

Pictures from the crime scene

Exchange admin center: Migration details

Exchange admin center: Migration details

Webmail log-in error

Webmail log-in error

Microsoft terms that I (still) have not accepted

Note that the closing button in the top right disappeared for me on April 15th.

Microsoft terms not accepted

DomainFactory customer satisfaction guarantee

Customer satisfaction guarantee