Is there a way to set a context limit for each connection not to take up all the CPU or MEM. We have users using Anlaytical view and fired of queries that didn't make sense and basically asked for all the data in HANA and which in turn took up all the CPU and MEM. There should be a way to control this.
Below is the query they ran basically asked for all the FI documents which there are about 1.3B records and no where clause. This brought HANA to it's Knees peaked the CPU to 97% for hours and hours and mem consumption extremly high almost 90% also.
SELECT
"_SYS_BIC"."fi-gl-lineitems/AN_BKPF_BSEG"."BUKRS",
"_SYS_BIC"."fi-gl-lineitems/AN_BKPF_BSEG"."HKONT_SUBSTR",
Cast ( TO_DATE("_SYS_BIC"."fi-gl-lineitems/AN_BKPF_BSEG"."AUGDT", 'YYYYMMDD') as DATE),
"_SYS_BIC"."fi-gl-lineitems/AN_BKPF_BSEG"."AUGBL",
Cast ( TO_DATE("_SYS_BIC"."fi-gl-lineitems/AN_BKPF_BSEG"."BUDAT", 'YYYYMMDD') as DATE),
Cast ( TO_DATE("_SYS_BIC"."fi-gl-lineitems/AN_BKPF_BSEG"."BLDAT", 'YYYYMMDD') as DATE),
"_SYS_BIC"."fi-gl-lineitems/AN_BKPF_BSEG"."BELNR",
"_SYS_BIC"."fi-gl-lineitems/AN_BKPF_BSEG"."BLART",
"_SYS_BIC"."fi-gl-lineitems/AN_BKPF_BSEG"."XBLNR",
"_SYS_BIC"."fi-gl-lineitems/AN_BKPF_BSEG"."BKTXT",
"_SYS_BIC"."fi-gl-lineitems/AN_BKPF_BSEG"."BUZEI",
sum("_SYS_BIC"."fi-gl-lineitems/AN_BKPF_BSEG"."DMBTR_NEW")
FROM
"_SYS_BIC"."fi-gl-lineitems/AN_BKPF_BSEG"
GROUP BY
"_SYS_BIC"."fi-gl-lineitems/AN_BKPF_BSEG"."BUKRS",
"_SYS_BIC"."fi-gl-lineitems/AN_BKPF_BSEG"."HKONT_SUBSTR",
Cast ( TO_DATE("_SYS_BIC"."fi-gl-lineitems/AN_BKPF_BSEG"."AUGDT", 'YYYYMMDD') as DATE),
"_SYS_BIC"."fi-gl-lineitems/AN_BKPF_BSEG"."AUGBL",
Cast ( TO_DATE("_SYS_BIC"."fi-gl-lineitems/AN_BKPF_BSEG"."BUDAT", 'YYYYMMDD') as DATE),
Cast ( TO_DATE("_SYS_BIC"."fi-gl-lineitems/AN_BKPF_BSEG"."BLDAT", 'YYYYMMDD') as DATE),
"_SYS_BIC"."fi-gl-lineitems/AN_BKPF_BSEG"."BELNR",
"_SYS_BIC"."fi-gl-lineitems/AN_BKPF_BSEG"."BLART",
"_SYS_BIC"."fi-gl-lineitems/AN_BKPF_BSEG"."XBLNR",
"_SYS_BIC"."fi-gl-lineitems/AN_BKPF_BSEG"."BKTXT",
"_SYS_BIC"."fi-gl-lineitems/AN_BKPF_BSEG"."BUZEI"