# Dynatrace Managed failover mechanism

**URL:** <https://community.dynatrace.com/t/dynatrace-managed-failover-mechanism/31030>\
**Category:** Dynatrace Managed\
**Tags:** installation, dynatrace-managed\
**Created:** [April 6, 2021, 10:49am UTC](https://community.dynatrace.com/t/dynatrace-managed-failover-mechanism/31030 "2021-04-06T10:49:38Z")\
**Posts on this page:** 6\
**Page:** 1

<div class="post-metadata">

**Author:** ![SB\_12](https://avatars.discourse-cdn.com/v4/letter/s/3ab097/32.png) [@SB\_12](https://community.dynatrace.com/u/SB_12)\
**Post date:** [April 6, 2021, 10:49am UTC](https://community.dynatrace.com/t/dynatrace-managed-failover-mechanism/31030/1 "2021-04-06T10:49:38Z")

</div>

Hi , can someone please help me understand&nbsp;Managed failover mechanism . the explanation in document is confusing .

[Managed failover mechanism](https://docs.dynatrace.com/managed/shortlink/managed-cluster-failover)

Is it correct to say that incase of a node failure , I will be able to have details of user session and metrics . However , transactional data will be lost for the time when node was down?

What is the replication factor of 3 means?

Thanks in advance.

---

<div class="post-metadata">

**Author:** ![RazTN7](https://dub1.discourse-cdn.com/dynatrace/user_avatar/community.dynatrace.com/raztn7/32/10208_2.png) [@RazTN7](https://community.dynatrace.com/u/RazTN7)\
**Post date:** [April 6, 2021, 12:26pm UTC](https://community.dynatrace.com/t/dynatrace-managed-failover-mechanism/31030/2 "2021-04-06T12:26:39Z")

</div>

Hello&nbsp;[@SB\_12](https://community.dynatrace.com/user/36508)

Quick summary - If you have 3node servers if say 1node is down. Still Dynatrace is capable to monitor with 2node servers.

Latency required is 10ms across 3node servers to sync between them. When a transaction is monitored its evenly distributed across 3nodes. So that no data are missed even if you have large volume of data coming to node servers.

Note- If their is high latency between Node servers. Then case can be upside down.

Cheers

R

---

<div class="post-metadata">

**Author:** ![SB\_12](https://avatars.discourse-cdn.com/v4/letter/s/3ab097/32.png) [@SB\_12](https://community.dynatrace.com/u/SB_12)\
**Post date:** [April 6, 2021, 12:43pm UTC](https://community.dynatrace.com/t/dynatrace-managed-failover-mechanism/31030/3 "2021-04-06T12:43:01Z")

</div>

hi , thanks for your reply. So that means that Transactional data will not be completely captured incase one of the nodes get down.

Metrics and user session will be available on either of the two nodes( which are up) .

---

<div class="post-metadata">

**Author:** ![RazTN7](https://dub1.discourse-cdn.com/dynatrace/user_avatar/community.dynatrace.com/raztn7/32/10208_2.png) [@RazTN7](https://community.dynatrace.com/u/RazTN7)\
**Post date:** [April 6, 2021, 12:46pm UTC](https://community.dynatrace.com/t/dynatrace-managed-failover-mechanism/31030/4 "2021-04-06T12:46:06Z")

</div>

Yes. In any case one of your node goes down. Still data is available on other node servers

Cheers!

R

---

<div class="post-metadata">

**Author:** ![RazTN7](https://dub1.discourse-cdn.com/dynatrace/user_avatar/community.dynatrace.com/raztn7/32/10208_2.png) [@RazTN7](https://community.dynatrace.com/u/RazTN7)\
**Post date:** [April 12, 2021, 3:13pm UTC](https://community.dynatrace.com/t/dynatrace-managed-failover-mechanism/31030/5 "2021-04-12T15:13:04Z")

</div>

[@SB\_12](https://community.dynatrace.com/user/36508)&nbsp;hope this helped.

---

<div class="post-metadata">

**Author:** ![Marcelo\_santand](https://dub1.discourse-cdn.com/dynatrace/user_avatar/community.dynatrace.com/marcelo_santand/32/13828_2.png) [@Marcelo\_santand](https://community.dynatrace.com/u/Marcelo_santand)\
**Post date:** [April 21, 2021, 7:08pm UTC](https://community.dynatrace.com/t/dynatrace-managed-failover-mechanism/31030/6 "2021-04-21T19:08:22Z")

</div>

He,

Ideally for critical environments you should have at least 3 servers to work properly. We have managed and 3 servers, very recently one of the disks of 1 server was corrupted and we had to repair it together with the system administrators. If we did not have 3 nodes and only 1, we would lose all the data. 100% recommended to have 3 nodes. NOTE, latency is important as well as time synchronization or time zone, you must use NTP.

regard!
