# How does the service link work

**URL:** <https://forums.suse.com/t/how-does-the-service-link-work/1830>\
**Category:** Rancher 1.x\
**Created:** [February 23, 2016, 12:44pm UTC](https://forums.suse.com/t/how-does-the-service-link-work/1830 "2016-02-23T12:44:48Z")\
**Posts on this page:** 3\
**Page:** 1

<div class="post-metadata">

**Author:** ![Johan\_Lindberg](https://avatars.discourse-cdn.com/v4/letter/j/6a8cbe/32.png) [@Johan\_Lindberg](https://forums.suse.com/u/Johan_Lindberg)\
**Post date:** [February 23, 2016, 12:44pm UTC](https://forums.suse.com/t/how-does-the-service-link-work/1830/1 "2016-02-23T12:44:48Z")

</div>

If i link my service to f.eks an mongodb cluster with three instances.  
How does the link determine which node to link to?

Will it work somewhat as a load balancer?  
Or does it simply give an dns result with multiple addresses?

---

<div class="post-metadata">

**Author:** ![vincent](https://sea2.discourse-cdn.com/flex022/user_avatar/forums.suse.com/vincent/32/7156_2.png) [@vincent](https://forums.suse.com/u/vincent)\
**Post date:** [February 23, 2016, 3:28pm UTC](https://forums.suse.com/t/how-does-the-service-link-work/1830/2 "2016-02-23T15:28:26Z")

</div>

It returns multiple A records in random order with short TTLs, which for some DNS clients works somewhat as a balancer (DNS round robin), and for others [not so much](https://github.com/rancher/rancher/issues/3495).

---

<div class="post-metadata">

**Author:** ![Johan\_Lindberg](https://avatars.discourse-cdn.com/v4/letter/j/6a8cbe/32.png) [@Johan\_Lindberg](https://forums.suse.com/u/Johan_Lindberg)\
**Post date:** [February 23, 2016, 5:49pm UTC](https://forums.suse.com/t/how-does-the-service-link-work/1830/3 "2016-02-23T17:49:06Z")

</div>

Thanks for the clarification.  
Sounds like it should work well for me, if the .net default dns client sorts the list I believe I can manually randomize it until your suggestion is implemented.