Kubernetes Pod Resources

Component	Instances	CPU req	CPU limit	RAM req (GB)	RAM limit (GB)
Automate	2	-	-	-	-
admin	2	4	8	6	6
dialogs	2	4	8	18	18
nlu-pipeline	2	1	4	2	2
nlu-facade	2	1	4	5	5
gateway	2	4	8	2	2
analytics	2	1	4	2	2
cron-orchestrator	2	1	4	2	2
web-chat	2	1	4	2	2
sso-server	2	1	4	2	2
storage	2	2	2	2	2
channels-connector*	2	1	4	2	2
thread-coordinator*	2	1	4	2	2
bot-integration	2	1	2	2	2
refinery	2	1	2	8	8
uploader	2	1	2	2	2

Listen&React*	2
new-web	2	1	2	4	4
analyser	2	1	2	7	7

Infrastructure	2	-	-	-	-
ElasticSearch	3	1	4	16	16
RabbitMQ	3	1	2	6	6
PostgreSQL	3	2	4	8	8
Kubernetes - master node	3	8	8	16	16
NFS	3	1	2	2	2
Redis AI	1	18	36	36	36

Voice systems*	2	-	-	-	-
Voice Gateway	2	48	48	768	768

NLU	2	-	-	-	-
duckling	2	2	4	1	1
inferrer	2	1	2	1	1
intentizer-multi	2	2	2	0.2	0.2
intentizer-multi-fitter	2	9	16	16	16
keywords	2	1	2	0.25	0.25
name-service	2	0.5	1	0.25	0.25
ner-pl	2	4	8	2	2
pattern	2	2	2	2	2
pcre	2	1	2	0.5	0.5
sentiduck	2	4	8	1	1
tokenizer-pl	2	3	6	2	2
tagger-de	2	2	2	2	2
tagger-en	2	2	2	2	2
tagger-multi	2	2	2	2	2
ner-multi	2	2	10	10	10

Resources summary

Component	Instances	CPU req	CPU limit	RAM req (GB)	RAM limit (GB)
Automate	2	50	128	112 (115**)	112 (115**)
Listen&React	2	4	8	22	22
Infrastructure	3	57	96	180	180
Voice systems	2	96	96	1536	1536
NLU	2	71	138	85	85
	TOTAL	278	466	1935	1935

Recommended VM configuration

	qty.	HDD [GB]	vCPU	RAM [GB]
Kubernetes Master	3	75	8	16
Kubernetes Worker	6	200	36	80
Voice Gateway*	2	250	48	768

* - Optional applications

⚠️ When planning deployment please take into account requirements described in Deployment Assumptions and contact us for further details.

ℹ️ RedisAI assumes 35 active NLU simple models. If more active NLU models are needed please calculate 0.8GB per 1 NLU simple model. If complex models are used, calculate 4GB per 1 model - more information about complex models in Comparison of available intentizer types